---
######################################################################
# File      :   pigsty.yml
# Desc      :   Pigsty Configuration Example
# Note      :   Pigsty Sandbox Demo
# Link      :   https://pigsty.cc/zh/docs/config/
# Ctime     :   2020-05-22
# Mtime     :   2021-04-19
# Copyright (C) 2018-2021 Ruohang Feng
######################################################################


######################################################################
#               Development Environment Inventory                    #
######################################################################


all: # top-level namespace

  #==================================================================#
  #                           Clusters                               #
  #==================================================================#
  # postgres database clusters are defined as kv pair in `all.children`
  # where the key is cluster name and the value is the object consist
  # of cluster members (hosts) and cluster specific variables (vars)
  # meta nodes are defined in special group "meta" with `meta_node=true`

  children:

    #-----------------------------
    # meta controller
    #-----------------------------
    meta:      # special group 'meta' defines the main controller machine
      vars:
        meta_node: true                    # mark node as meta controller
        ansible_group_priority: 99         # meta group has top priority
      hosts:
        10.10.10.10: {}

    #-----------------------------
    # cluster: pg-meta
    #-----------------------------
    # pg-meta is a single-node pgsql cluster deployed on meta node (10.10.10.10)
    pg-meta:
      # - cluster members - #
      hosts:
        10.10.10.10: {pg_seq: 1, pg_role: primary, pg_offline_query: true}

      # - cluster configs - #
      vars:
        pg_cluster: pg-meta                 # define actual cluster name
        pg_version: 13                      # define installed pgsql version
        node_tune: tiny                     # tune node into oltp|olap|crit|tiny mode
        pg_conf: tiny.yml                   # tune pgsql into oltp|olap|crit|tiny mode
        patroni_mode: pause                 # enter maintenance mode, {default|pause|remove}
        patroni_watchdog_mode: off          # disable watchdog (require|automatic|off)
        pg_lc_ctype: en_US.UTF8             # enabled pg_trgm i18n char support

        # - defining business users - #
        pg_users:
          # default production read-write user dbuser_meta
          - name: dbuser_meta                              # user's name is required
            password: md5d3d10d8cad606308bdb180148bf663e1  # md5 password is acceptable
            pgbouncer: true                                # add user to pgbouncer userlist
            roles: [dbrole_readwrite]                      # grant roles to user
            comment: default production read-write user for meta database

          # default production read-only user for grafana direct access
          - name: dbuser_grafana
            password: DBUser.Grafana
            pgbouncer: true
            roles: [dbrole_readonly]
            comment: default readonly access for grafana datasource

          # complete example of user/role definition
          - name: dbuser_pigsty             # pigsty user have admin access (DDL|DML)
            password: DBUser.Pigsty         # example user's password, can be md5 encrypted
            login: true                     # can login, true by default (should be false for role)
            superuser: false                # is superuser? false by default
            createdb: false                 # can create database? false by default
            createrole: false               # can create role? false by default
            inherit: true                   # can this role use inherited privileges?
            replication: false              # can this role do replication? false by default
            bypassrls: false                # can this role bypass row level security? false by default
            pgbouncer: true                 # add this user to pgbouncer? false by default (true for production user)
            connlimit: -1                   # connection limit, -1 disable limit
            expire_in: 3650                 # now + n days when this role is expired (OVERWRITE expire_at)
            expire_at: '2030-12-31'         # 'timestamp' when this role is expired (OVERWRITTEN by expire_in)
            comment: pigsty admin user      # comment on user/role
            roles: [dbrole_admin]           # dbrole_{admin,readonly,readwrite,offline}
            parameters:                     # additional role level parameters with ALTER ROLE SET
              search_path: pigsty,public    # add pigsty schema into search_path

        # - defining business databases - #
        pg_databases:
          - name: meta                      # name is the only required field for a database
            # baseline: metadb/schema.sql   # pigsty meta database baseline
            # owner: postgres               # optional, database owner
            # template: template1           # optional, template1 by default
            # encoding: UTF8                # optional, UTF8 by default , must same as template database, leave blank to set to db default
            # locale: C                     # optional, C by default , must same as template database, leave blank to set to db default
            # lc_collate: C                 # optional, C by default , must same as template database, leave blank to set to db default
            # lc_ctype: C                   # optional, C by default , must same as template database, leave blank to set to db default
            # tablespace: pg_default        # optional, 'pg_default' is the default tablespace
            # allowconn: true               # optional, true by default, false disable connect at all
            # revokeconn: false             # optional, false by default, true revoke connect from public # (only default user and owner have connect privilege on database)
            # pgbouncer: true               # optional, add this database to pgbouncer list? true by default
            comment: pigsty meta database   # optional, comment string for database
            connlimit: -1                   # optional, connection limit, -1 or none disable limit (default)
            schemas: [pigsty]               # optional, create additional schema
            extensions:                     # optional, extension name and which schema to create
              - {name: adminpack, schema: pg_catalog}
            parameters:                       # optional, extra parameters with ALTER DATABASE
              search_path: 'pigsty,public'    # add pigsty to search_path

        pg_default_database: meta           # default database will be used as primary monitor target
        vip_mode: l2                        # none|l2|l4, l2 vip are used in sandbox demo
        vip_address: 10.10.10.2             # virtual ip address
        vip_cidrmask: 8                     # cidr network mask length
        vip_interface: eth1                 # interface to add virtual ip


    #-----------------------------
    # cluster: pg-test
    #-----------------------------
    # uncomment this for complete 4-node sandbox demo environment

    #pg-test: # define cluster named 'pg-test'
    #  # - cluster members - #
    #  hosts:
    #    10.10.10.11: {pg_seq: 1, pg_role: primary}
    #    10.10.10.12: {pg_seq: 2, pg_role: replica}
    #    10.10.10.13: {pg_seq: 3, pg_role: offline}
    #
    #  # - cluster configs - #
    #  vars:
    #    # basic settings
    #    pg_cluster: pg-test                 # define actual cluster name
    #    pg_version: 13                      # define installed pgsql version
    #    node_tune: tiny                     # tune node into oltp|olap|crit|tiny mode
    #    pg_conf: tiny.yml                   # tune pgsql into oltp|olap|crit|tiny mode
    #    pg_users:
    #      - name: test                      # admin user for pg-test, have DDL
    #        password: test
    #        roles: [dbrole_admin]
    #        pgbouncer: true
    #        comment: default admin user for test database
    #
    #      - name: dbuser_test               # production rw-user
    #        password: DBUser.Test
    #        roles: [dbrole_readwrite]
    #        pgbouncer: true
    #        comment: default test user for production usage
    #
    #    pg_databases:                       # create a business database 'test'
    #      - name: test                      # use the simplest form
    #        extensions:                     # install postgis to test database
    #          - {name: postgis, schema: public}
    #    pg_default_database: test           # default database will be used as primary monitor target
    #
    #    # extra service settings
    #    pg_services_extra:                  # extra services to be added
    #      - name: standby                   # service name pg-meta-standby
    #        src_ip: "*"
    #        src_port: 5435                  # 5435 routes to sync replica
    #        dst_port: postgres
    #        check_url: /sync                # use /sync health check
    #        selector: "[]"                  # jmespath to filter instances
    #        selector_backup: "[? pg_role == `primary`]"  # primary used as backup server for standby service
    #
    #    # proxy settings
    #    vip_mode: l2                        # enable/disable vip (require members in same LAN)
    #    vip_address: 10.10.10.3             # virtual ip address
    #    vip_cidrmask: 8                     # cidr network mask length
    #    vip_interface: eth1                 # interface to add virtual ip


  #==================================================================#
  #                           Globals                                #
  #==================================================================#
  vars:

    #------------------------------------------------------------------------------
    # CONNECTION PARAMETERS
    #------------------------------------------------------------------------------
    # this section defines connection parameters

    # ansible_user: vagrant                       # admin user with ssh access and sudo privilege

    proxy_env: # global proxy env when downloading packages
      no_proxy: "localhost,127.0.0.1,10.0.0.0/8,192.168.0.0/16,*.pigsty,*.aliyun.com,mirrors.aliyuncs.com,mirrors.tuna.tsinghua.edu.cn,mirrors.zju.edu.cn,*.myqcloud.com"
      # http_proxy: ''
      # https_proxy: ''
      # all_proxy: ''


    #------------------------------------------------------------------------------
    # REPO PROVISION
    #------------------------------------------------------------------------------
    # this section defines how to build a local repo

    # - repo basic - #
    repo_enabled: true                            # build local yum repo on meta nodes?
    repo_name: pigsty                             # local repo name
    repo_address: yum.pigsty                      # repo external address (ip:port or url)
    repo_port: 80                                 # listen address, must same as repo_address
    repo_home: /www                               # default repo dir location
    repo_rebuild: false                           # force re-download packages
    repo_remove: true                             # remove existing repos

    # - where to download - #
    repo_upstreams:
      - name: base
        description: CentOS-$releasever - Base - Aliyun Mirror
        baseurl:
          - http://mirrors.aliyun.com/centos/$releasever/os/$basearch/
          - http://mirrors.aliyuncs.com/centos/$releasever/os/$basearch/
          - http://mirrors.cloud.aliyuncs.com/centos/$releasever/os/$basearch/
        gpgcheck: no
        failovermethod: priority

      - name: updates
        description: CentOS-$releasever - Updates - Aliyun Mirror
        baseurl:
          - http://mirrors.aliyun.com/centos/$releasever/updates/$basearch/
          - http://mirrors.aliyuncs.com/centos/$releasever/updates/$basearch/
          - http://mirrors.cloud.aliyuncs.com/centos/$releasever/updates/$basearch/
        gpgcheck: no
        failovermethod: priority

      - name: extras
        description: CentOS-$releasever - Extras - Aliyun Mirror
        baseurl:
          - http://mirrors.aliyun.com/centos/$releasever/extras/$basearch/
          - http://mirrors.aliyuncs.com/centos/$releasever/extras/$basearch/
          - http://mirrors.cloud.aliyuncs.com/centos/$releasever/extras/$basearch/
        gpgcheck: no
        failovermethod: priority

      - name: epel
        description: CentOS $releasever - EPEL - Aliyun Mirror
        baseurl: http://mirrors.aliyun.com/epel/$releasever/$basearch
        gpgcheck: no
        failovermethod: priority

      - name: grafana
        description: Grafana - TsingHua Mirror
        gpgcheck: no
        baseurl: https://mirrors.tuna.tsinghua.edu.cn/grafana/yum/rpm

      - name: prometheus
        description: Prometheus and exporters
        gpgcheck: no
        baseurl: https://packagecloud.io/prometheus-rpm/release/el/$releasever/$basearch

      # consider using ZJU PostgreSQL mirror in mainland china
      - name: pgdg-common
        description: PostgreSQL common RPMs for RHEL/CentOS $releasever - $basearch
        gpgcheck: no
        baseurl: https://download.postgresql.org/pub/repos/yum/common/redhat/rhel-$releasever-$basearch
        # baseurl: http://mirrors.zju.edu.cn/postgresql/repos/yum/common/redhat/rhel-$releasever-$basearch

      - name: pgdg13
        description: PostgreSQL 13 for RHEL/CentOS $releasever - $basearch
        gpgcheck: no
        baseurl: https://download.postgresql.org/pub/repos/yum/13/redhat/rhel-$releasever-$basearch
        # baseurl: http://mirrors.zju.edu.cn/postgresql/repos/yum/13/redhat/rhel-$releasever-$basearch

      - name: centos-sclo
        description: CentOS-$releasever - SCLo
        gpgcheck: no
        mirrorlist: http://mirrorlist.centos.org?arch=$basearch&release=7&repo=sclo-sclo

      - name: centos-sclo-rh
        description: CentOS-$releasever - SCLo rh
        gpgcheck: no
        mirrorlist: http://mirrorlist.centos.org?arch=$basearch&release=7&repo=sclo-rh

      - name: nginx
        description: Nginx Official Yum Repo
        skip_if_unavailable: true
        gpgcheck: no
        baseurl: http://nginx.org/packages/centos/$releasever/$basearch/

      - name: haproxy
        description: Copr repo for haproxy
        skip_if_unavailable: true
        gpgcheck: no
        baseurl: https://download.copr.fedorainfracloud.org/results/roidelapluie/haproxy/epel-$releasever-$basearch/

      # for latest consul & kubernetes
      - name: harbottle
        description: Copr repo for main owned by harbottle
        skip_if_unavailable: true
        gpgcheck: no
        baseurl: https://download.copr.fedorainfracloud.org/results/harbottle/main/epel-$releasever-$basearch/

    # - what to download - #
    repo_packages:
      # repo bootstrap packages
      - epel-release nginx wget yum-utils yum createrepo sshpass unzip                        # bootstrap packages

      # node basic packages
      - ntp chrony uuid lz4 nc pv jq vim-enhanced make patch bash lsof wget git tuned         # basic system util
      - readline zlib openssl libyaml libxml2 libxslt perl-ExtUtils-Embed ca-certificates     # basic pg dependency
      - numactl grubby sysstat dstat iotop bind-utils net-tools tcpdump socat ipvsadm telnet  # system utils

      # dcs & monitor packages
      - grafana prometheus2 pushgateway alertmanager                                          # monitor and ui
      - node_exporter postgres_exporter nginx_exporter blackbox_exporter                      # exporter
      - consul consul_exporter consul-template etcd                                           # dcs

      # python3 dependencies
      - ansible python python-pip python-psycopg2 audit                                       # ansible & python
      - python3 python3-psycopg2 python36-requests python3-etcd python3-consul                # python3
      - python36-urllib3 python36-idna python36-pyOpenSSL python36-cryptography               # patroni extra deps

      # proxy and load balancer
      - haproxy keepalived dnsmasq                                                            # proxy and dns

      # postgres common Packages
      - patroni patroni-consul patroni-etcd pgbouncer pg_cli pgbadger pg_activity               # major components
      - pgcenter boxinfo check_postgres emaj pgbconsole pg_bloat_check pgquarrel                # other common utils
      - barman barman-cli pgloader pgFormatter pitrery pspg pgxnclient PyGreSQL pgadmin4 tail_n_mail

      # postgres 13 packages
      - postgresql13* postgis31* citus_13 timescaledb_13 # pgrouting_13                         # postgres 13 and postgis 31
      - pg_repack13 pg_squeeze13                                                                # maintenance extensions
      - pg_qualstats13 pg_stat_kcache13 system_stats_13 bgw_replstatus13                        # stats extensions
      - plr13 plsh13 plpgsql_check_13 plproxy13 plr13 plsh13 plpgsql_check_13 pldebugger13      # PL extensions
      - hdfs_fdw_13 mongo_fdw13 mysql_fdw_13 ogr_fdw13 redis_fdw_13 pgbouncer_fdw13             # FDW extensions
      - wal2json13 count_distinct13 ddlx_13 geoip13 orafce13                                    # MISC extensions
      - rum_13 hypopg_13 ip4r13 jsquery_13 logerrors_13 periods_13 pg_auto_failover_13 pg_catcheck13
      - pg_fkpart13 pg_jobmon13 pg_partman13 pg_prioritize_13 pg_track_settings13 pgaudit15_13
      - pgcryptokey13 pgexportdoc13 pgimportdoc13 pgmemcache-13 pgmp13 pgq-13
      - pguint13 pguri13 prefix13  safeupdate_13 semver13  table_version13 tdigest13

    repo_url_packages:
      # additional rpm packages
      - https://github.com/Vonng/pg_exporter/releases/download/v0.3.2/pg_exporter-0.3.2-1.el7.x86_64.rpm
      - https://github.com/cybertec-postgresql/vip-manager/releases/download/v0.6/vip-manager_0.6-1_amd64.rpm
      - http://guichaz.free.fr/polysh/files/polysh-0.4-1.noarch.rpm

      # tar.gz and zip binary packages
      - https://github.com/prometheus/node_exporter/releases/download/v1.1.2/node_exporter-1.1.2.linux-amd64.tar.gz # monitor binary
      - https://github.com/Vonng/pg_exporter/releases/download/v0.3.2/pg_exporter_v0.3.2_linux-amd64.tar.gz
      - https://github.com/grafana/loki/releases/download/v2.2.1/loki-linux-amd64.zip           # loki binary
      - https://github.com/grafana/loki/releases/download/v2.2.1/promtail-linux-amd64.zip
      - https://github.com/grafana/loki/releases/download/v2.2.1/logcli-linux-amd64.zip
      - https://github.com/grafana/loki/releases/download/v2.2.1/loki-canary-linux-amd64.zip

      # mirror in mainland china (use commented packages to install from official site)
      # - http://pigsty-1304147732.cos.accelerate.myqcloud.com/pkg/pg_exporter-0.3.2-1.el7.x86_64.rpm
      # - http://pigsty-1304147732.cos.accelerate.myqcloud.com/pkg/vip-manager_0.6-1_amd64.rpm
      # - http://pigsty-1304147732.cos.accelerate.myqcloud.com/pkg/polysh-0.4-1.noarch.rpm

    #------------------------------------------------------------------------------
    # NODE PROVISION
    #------------------------------------------------------------------------------
    # this section defines how to provision nodes
    # nodename:                                   # if defined, node's hostname will be overwritten

    # - node dns - #
    node_dns_hosts: # static dns records in /etc/hosts
      - 10.10.10.10 yum.pigsty
    node_dns_server: add                          # add (default) | none (skip) | overwrite (remove old settings)
    node_dns_servers:                             # dynamic nameserver in /etc/resolv.conf
      - 10.10.10.10
    node_dns_options:                             # dns resolv options
      - options single-request-reopen timeout:1 rotate
      - domain service.consul

    # - node repo - #
    node_repo_method: local                       # none|local|public (use local repo for production env)
    node_repo_remove: true                        # whether remove existing repo
    node_local_repo_url:                          # local repo url (if method=local, make sure firewall is configured or disabled)
      - http://yum.pigsty/pigsty.repo

    # - node packages - #
    node_packages:                                # common packages for all nodes
      - wget,yum-utils,sshpass,ntp,chrony,tuned,uuid,lz4,vim-minimal,make,patch,bash,lsof,wget,unzip,git,readline,zlib,openssl
      - numactl,grubby,sysstat,dstat,iotop,bind-utils,net-tools,tcpdump,socat,ipvsadm,telnet,tuned,pv,jq
      - python3,python3-psycopg2,python36-requests,python3-etcd,python3-consul
      - python36-urllib3,python36-idna,python36-pyOpenSSL,python36-cryptography
      - node_exporter,consul,consul-template,etcd,haproxy,keepalived,vip-manager
    node_extra_packages:                          # extra packages for all nodes
      - patroni,patroni-consul,patroni-etcd,pgbouncer,pgbadger,pg_activity
    node_meta_packages:                           # packages for meta nodes only
      - grafana,prometheus2,alertmanager,nginx_exporter,blackbox_exporter,pushgateway
      - dnsmasq,nginx,ansible,pgbadger,polysh

    # build & devel packages (add to repo_packages too if you want build database & extensions from source)
    # - gcc,gcc-c++,clang,coreutils,diffutils,rpm-build,rpm-devel,rpmlint,rpmdevtools
    # - zlib-devel,openssl-libs,openssl-devel,pam-devel,libxml2-devel,libxslt-devel,openldap-devel,systemd-devel,tcl-devel,python-devel


    # - node features - #
    node_disable_numa: false                      # disable numa, important for production database, reboot required
    node_disable_swap: false                      # disable swap, important for production database
    node_disable_firewall: true                   # disable firewall (required if using kubernetes)
    node_disable_selinux: true                    # disable selinux  (required if using kubernetes)
    node_static_network: true                     # keep dns resolver settings after reboot
    node_disk_prefetch: false                     # setup disk prefetch on HDD to increase performance

    # - node kernel modules - #
    node_kernel_modules:
      - softdog
      - br_netfilter
      - ip_vs
      - ip_vs_rr
      - ip_vs_rr
      - ip_vs_wrr
      - ip_vs_sh
      - nf_conntrack_ipv4

    # - node tuned - #
    node_tune: tiny                               # install and activate tuned profile: none|oltp|olap|crit|tiny
    node_sysctl_params: {}                        # set additional sysctl parameters, k:v format
    # net.bridge.bridge-nf-call-iptables: 1     # example kv parameters

    # - node user - #
    node_admin_setup: true                        # setup an default admin user ?
    node_admin_uid: 88                            # uid and gid for admin user
    node_admin_username: dba                      # default admin user: dba
    node_admin_ssh_exchange: true                 # exchange admin's ssh key among cluster ?
    node_admin_pk_current: false                  # add current user's ~/.ssh/id_rsa.pub to admin pk
    node_admin_pks:                               # ssh public keys to be added to admin user
      - 'ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAAAgQC7IMAMNavYtWwzAJajKqwdn3ar5BhvcwCnBTxxEkXhGlCO2vfgosSAQMEflfgvkiI5nM1HIFQ8KINlx1XLO7SdL5KdInG5LIJjAFh0pujS4kNCT9a5IGvSq1BrzGqhbEcwWYdju1ZPYBcJm/MG+JD0dYCh8vfrYB/cYMD0SOmNkQ== vagrant@pigsty.com'

    # - node ntp - #
    node_ntp_service: ntp                         # ntp or chrony
    node_ntp_config: true                         # overwrite existing ntp config?
    node_timezone: Asia/Shanghai                  # default node timezone
    node_ntp_servers:                             # default NTP servers
      - pool cn.pool.ntp.org iburst
      - pool pool.ntp.org iburst
      - pool time.pool.aliyun.com iburst
      - server 10.10.10.10 iburst


    #------------------------------------------------------------------------------
    # META PROVISION
    #------------------------------------------------------------------------------
    # - ca - #
    ca_method: create                             # create|copy|recreate
    ca_subject: "/CN=root-ca"                     # self-signed CA subject
    ca_homedir: /ca                               # ca cert directory
    ca_cert: ca.crt                               # ca public key/cert
    ca_key: ca.key                                # ca private key

    # - nginx - #
    nginx_upstream:
      - { name: home,          host: pigsty,   url: "127.0.0.1:3000"}
      - { name: consul,        host: c.pigsty, url: "127.0.0.1:8500" }
      - { name: grafana,       host: g.pigsty, url: "127.0.0.1:3000" }
      - { name: prometheus,    host: p.pigsty, url: "127.0.0.1:9090" }
      - { name: alertmanager,  host: a.pigsty, url: "127.0.0.1:9093" }
      - { name: haproxy,       host: h.pigsty, url: "127.0.0.1:9091" }

    # - nameserver - #
    dns_records: # dynamic dns record resolved by dnsmasq
      - 10.10.10.2  pg-meta                       # sandbox vip for pg-meta
      - 10.10.10.3  pg-test                       # sandbox vip for pg-test
      - 10.10.10.10 meta-1                        # sandbox node meta-1 (node-0)
      - 10.10.10.11 node-1                        # sandbox node node-1
      - 10.10.10.12 node-2                        # sandbox node node-2
      - 10.10.10.13 node-3                        # sandbox node node-3
      - 10.10.10.10 pigsty
      - 10.10.10.10 y.pigsty yum.pigsty
      - 10.10.10.10 c.pigsty consul.pigsty
      - 10.10.10.10 g.pigsty grafana.pigsty
      - 10.10.10.10 p.pigsty prometheus.pigsty
      - 10.10.10.10 a.pigsty alertmanager.pigsty
      - 10.10.10.10 n.pigsty ntp.pigsty
      - 10.10.10.10 h.pigsty haproxy.pigsty

    # - prometheus - #
    prometheus_data_dir: /export/prometheus/data  # prometheus data dir
    prometheus_options: '--storage.tsdb.retention=30d'
    prometheus_reload: false                      # reload prometheus instead of recreate it
    prometheus_sd_method: consul                  # service discovery method: static|consul|etcd
    prometheus_scrape_interval: 5s                # global scrape & evaluation interval
    prometheus_scrape_timeout: 4s                 # scrape timeout
    prometheus_sd_interval: 5s                    # service discovery refresh interval

    # - grafana - #
    grafana_url: http://admin:admin@10.10.10.10:3000 # grafana url
    grafana_admin_password: admin                    # default grafana admin user password
    grafana_plugin: install                          # none|install|reinstall
    grafana_cache: /www/pigsty/grafana/plugins.tgz   # path to grafana plugins tarball
    grafana_customize: false                         # customize grafana resources
    grafana_plugins: # default grafana plugins list
      - redis-datasource
      - simpod-json-datasource
      - fifemon-graphql-datasource
      - sbueringer-consul-datasource
      - camptocamp-prometheus-alertmanager-datasource
      - ryantxu-ajax-panel
      - marcusolsson-hourly-heatmap-panel
      - michaeldmoore-multistat-panel
      - marcusolsson-treemap-panel
      - pr0ps-trackmap-panel
      - dalvany-image-panel
      - magnesium-wordcloud-panel
      - cloudspout-button-panel
      - speakyourcode-button-panel
      - jdbranham-diagram-panel
      - grafana-piechart-panel
      - snuids-radar-panel
      - digrich-bubblechart-panel
    grafana_git_plugins:
      - https://github.com/Vonng/grafana-echarts

    # - loki - #
    loki_clean: false                 # whether remove existing loki data
    loki_data_dir: /export/loki       # default loki data dir


    #------------------------------------------------------------------------------
    # DCS PROVISION
    #------------------------------------------------------------------------------
    service_registry: consul                      # where to register services: none | consul | etcd | both
    dcs_type: consul                              # consul | etcd | both
    dcs_name: pigsty                              # consul dc name | etcd initial cluster token
    dcs_servers:                                  # dcs server dict in name:ip format
      meta-1: 10.10.10.10                         # you could use existing dcs cluster
      # meta-2: 10.10.10.11                       # host which have their IP listed here will be init as server
      # meta-3: 10.10.10.12                       # 3 or 5 dcs nodes are recommend for production environment
    dcs_exists_action: clean                      # abort|skip|clean if dcs server already exists
    dcs_disable_purge: false                      # set to true to disable purge functionality for good (force dcs_exists_action = abort)
    consul_data_dir: /var/lib/consul              # consul data dir (/var/lib/consul by default)
    etcd_data_dir: /var/lib/etcd                  # etcd data dir (/var/lib/consul by default)


    #------------------------------------------------------------------------------
    # POSTGRES INSTALLATION
    #------------------------------------------------------------------------------
    # - dbsu - #
    pg_dbsu: postgres                             # os user for database, postgres by default (change it is not recommended!)
    pg_dbsu_uid: 26                               # os dbsu uid and gid, 26 for default postgres users and groups
    pg_dbsu_sudo: limit                           # none|limit|all|nopass (Privilege for dbsu, limit is recommended)
    pg_dbsu_home: /var/lib/pgsql                  # postgresql binary
    pg_dbsu_ssh_exchange: false                   # exchange ssh key among same cluster

    # - postgres packages - #
    pg_version: 13                                # default postgresql version
    pgdg_repo: false                              # use official pgdg yum repo (disable if you have local mirror)
    pg_add_repo: false                            # add postgres related repo before install (useful if you want a simple install)
    pg_bin_dir: /usr/pgsql/bin                    # postgres binary dir
    pg_packages:
      - postgresql${pg_version}*
      - postgis31_${pg_version}*
      - pgbouncer patroni pg_exporter pgbadger
      - patroni patroni-consul patroni-etcd pgbouncer pgbadger pg_activity
      - python3 python3-psycopg2 python36-requests python3-etcd python3-consul
      - python36-urllib3 python36-idna python36-pyOpenSSL python36-cryptography

    pg_extensions:
      - pg_repack${pg_version} pg_qualstats${pg_version} pg_stat_kcache${pg_version} wal2json${pg_version}
      # - ogr_fdw${pg_version} mysql_fdw_${pg_version} redis_fdw_${pg_version} mongo_fdw${pg_version} hdfs_fdw_${pg_version}
      # - count_distinct${version}  ddlx_${version}  geoip${version}  orafce${version}                                   # popular features
      # - hypopg_${version}  ip4r${version}  jsquery_${version}  logerrors_${version}  periods_${version}  pg_auto_failover_${version}  pg_catcheck${version}
      # - pg_fkpart${version}  pg_jobmon${version}  pg_partman${version}  pg_prioritize_${version}  pg_track_settings${version}  pgaudit15_${version}
      # - pgcryptokey${version}  pgexportdoc${version}  pgimportdoc${version}  pgmemcache-${version}  pgmp${version}  pgq-${version}  pgquarrel pgrouting_${version}
      # - pguint${version}  pguri${version}  prefix${version}   safeupdate_${version}  semver${version}   table_version${version}  tdigest${version}



    #------------------------------------------------------------------------------
    # POSTGRES PROVISION
    #------------------------------------------------------------------------------
    # - identity - #
    # pg_cluster:                                 # [REQUIRED] cluster name (cluster level,  validated during pg_preflight)
    # pg_seq: 0                                   # [REQUIRED] instance seq (instance level, validated during pg_preflight)
    # pg_role: replica                            # [REQUIRED] service role (instance level, validated during pg_preflight)
    # pg_shard:                                   # [OPTIONAL] shard name  (cluster level)
    # pg_sindex:                                  # [OPTIONAl] shard index (cluster level)

    # - identity option -#
    pg_hostname: false                            # overwrite node hostname with pg instance name
    pg_nodename: true                             # overwrite consul nodename with pg instance name

    # - retention - #
    # pg_exists_action, available options: abort|clean|skip
    #  - abort: abort entire play's execution (default)
    #  - clean: remove existing cluster (dangerous)
    #  - skip: end current play for this host
    # pg_exists: false                            # auxiliary flag variable (DO NOT SET THIS)
    pg_exists_action: clean
    pg_disable_purge: false                       # set to true to disable pg purge functionality for good (force pg_exists_action = abort)

    # - storage - #
    pg_data: /pg/data                             # postgres data directory
    pg_fs_main: /export                           # data disk mount point     /pg -> {{ pg_fs_main }}/postgres/{{ pg_instance }}
    pg_fs_bkup: /var/backups                      # backup disk mount point   /pg/* -> {{ pg_fs_bkup }}/postgres/{{ pg_instance }}/*

    # - connection - #
    pg_listen: '0.0.0.0'                          # postgres listen address, '0.0.0.0' by default (all ipv4 addr)
    pg_port: 5432                                 # postgres port (5432 by default)
    pg_localhost: /var/run/postgresql             # localhost unix socket dir for connection
    # pg_upstream:                                # [OPTIONAL] specify replication upstream (set on primary transform cluster into a standby cluster)


    # - patroni - #
    # patroni_mode, available options: default|pause|remove
    #   - default: default ha mode
    #   - pause:   into maintenance mode
    #   - remove:  remove patroni after bootstrap
    patroni_mode: default                         # pause|default|remove
    pg_namespace: /pg                             # top level key namespace in dcs
    patroni_port: 8008                            # default patroni port
    patroni_watchdog_mode: automatic              # watchdog mode: off|automatic|required
    pg_conf: tiny.yml                             # user provided patroni config template path

    # - flags - #
    pg_backup: false                              # store base backup on this node
    pg_delay: 0                                   # apply delay for offline|delayed instance

    # - localization - #
    pg_encoding: UTF8                             # default to UTF8
    pg_locale: C                                  # default to C
    pg_lc_collate: C                              # default to C
    pg_lc_ctype: en_US.UTF8                       # default to en_US.UTF8

    # - pgbouncer - #
    pgbouncer_port: 6432                          # pgbouncer port (6432 by default)
    pgbouncer_poolmode: transaction               # pooling mode: (transaction pooling by default)
    pgbouncer_max_db_conn: 100                    # important! do not set this larger than postgres max conn or conn limit


    #------------------------------------------------------------------------------
    # POSTGRES TEMPLATE
    #------------------------------------------------------------------------------
    # - template - #
    pg_init: pg-init                              # init script for cluster template

    # - system roles - #
    pg_replication_username: replicator           # system replication user
    pg_replication_password: DBUser.Replicator    # system replication password
    pg_monitor_username: dbuser_monitor           # system monitor user
    pg_monitor_password: DBUser.Monitor           # system monitor password
    pg_admin_username: dbuser_dba                 # system admin user
    pg_admin_password: DBUser.DBA                 # system admin password

    # - default roles - #
    # chekc http://pigsty.cc/zh/docs/concepts/provision/acl/ for more detail
    pg_default_roles:

      # common production readonly user
      - name: dbrole_readonly                 # production read-only roles
        login: false
        comment: role for global readonly access

      # common production read-write user
      - name: dbrole_readwrite                # production read-write roles
        login: false
        roles: [dbrole_readonly]             # read-write includes read-only access
        comment: role for global read-write access

      # offline have same privileges as readonly, but with limited hba access on offline instance only
      # for the purpose of running slow queries, interactive queries and perform ETL tasks
      - name: dbrole_offline
        login: false
        comment: role for restricted read-only access (offline instance)

      # admin have the privileges to issue DDL changes
      - name: dbrole_admin
        login: false
        bypassrls: true
        comment: role for object creation
        roles: [dbrole_readwrite,pg_monitor,pg_signal_backend]

      # dbsu, name is designated by `pg_dbsu`. It's not recommend to set password for dbsu
      - name: postgres
        superuser: true
        comment: system superuser

      # default replication user, name is designated by `pg_replication_username`, and password is set by `pg_replication_password`
      - name: replicator
        replication: true                          # for replication user
        bypassrls: true                            # logical replication require bypassrls
        roles: [pg_monitor, dbrole_readonly]       # logical replication require select privileges
        comment: system replicator

      # default monitor user, name is designated by `pg_monitor_username`, and password is set by `pg_monitor_password`
      - name: dbuser_monitor
        connlimit: 16
        comment: system monitor user
        roles: [pg_monitor, dbrole_readonly]
        parameters:
          log_min_duration_statement: 1000

      # default admin super user, name is designated by `pg_admin_username`, and password is set by `pg_admin_password`
      - name: dbuser_dba
        superuser: true
        comment: system admin user
        roles: [dbrole_admin]

      # default stats user, for ETL and slow queries
      - name: dbuser_stats
        password: DBUser.Stats
        comment: business offline user for offline queries and ETL
        roles: [dbrole_offline]


    # - privileges - #
    # object created by dbsu and admin will have their privileges properly set
    pg_default_privileges:
      - GRANT USAGE                         ON SCHEMAS   TO dbrole_readonly
      - GRANT SELECT                        ON TABLES    TO dbrole_readonly
      - GRANT SELECT                        ON SEQUENCES TO dbrole_readonly
      - GRANT EXECUTE                       ON FUNCTIONS TO dbrole_readonly
      - GRANT USAGE                         ON SCHEMAS   TO dbrole_offline
      - GRANT SELECT                        ON TABLES    TO dbrole_offline
      - GRANT SELECT                        ON SEQUENCES TO dbrole_offline
      - GRANT EXECUTE                       ON FUNCTIONS TO dbrole_offline
      - GRANT INSERT, UPDATE, DELETE        ON TABLES    TO dbrole_readwrite
      - GRANT USAGE,  UPDATE                ON SEQUENCES TO dbrole_readwrite
      - GRANT TRUNCATE, REFERENCES, TRIGGER ON TABLES    TO dbrole_admin
      - GRANT CREATE                        ON SCHEMAS   TO dbrole_admin

    # - schemas - #
    pg_default_schemas: [monitor]                 # default schemas to be created

    # - extension - #
    pg_default_extensions:                        # default extensions to be created
      - { name: 'pg_stat_statements',  schema: 'monitor' }
      - { name: 'pgstattuple',         schema: 'monitor' }
      - { name: 'pg_qualstats',        schema: 'monitor' }
      - { name: 'pg_buffercache',      schema: 'monitor' }
      - { name: 'pageinspect',         schema: 'monitor' }
      - { name: 'pg_prewarm',          schema: 'monitor' }
      - { name: 'pg_visibility',       schema: 'monitor' }
      - { name: 'pg_freespacemap',     schema: 'monitor' }
      - { name: 'pg_repack',           schema: 'monitor' }
      - name: postgres_fdw
      - name: file_fdw
      - name: btree_gist
      - name: btree_gin
      - name: pg_trgm
      - name: intagg
      - name: intarray

    # - hba - #
    pg_offline_query: false                       # set to true to enable offline query on instance
    pg_reload: true                               # reload postgres after hba changes
    pg_hba_rules:                                 # postgres host-based authentication rules
      - title: allow meta node password access
        role: common
        rules:
          - host    all     all                         10.10.10.10/32      md5

      - title: allow intranet admin password access
        role: common
        rules:
          - host    all     +dbrole_admin               10.0.0.0/8          md5
          - host    all     +dbrole_admin               172.16.0.0/12       md5
          - host    all     +dbrole_admin               192.168.0.0/16      md5

      - title: allow intranet password access
        role: common
        rules:
          - host    all             all                 10.0.0.0/8          md5
          - host    all             all                 172.16.0.0/12       md5
          - host    all             all                 192.168.0.0/16      md5

      - title: allow local read/write (local production user via pgbouncer)
        role: common
        rules:
          - local   all     +dbrole_readonly                                md5
          - host    all     +dbrole_readonly           127.0.0.1/32         md5

      - title: allow offline query (ETL,SAGA,Interactive) on offline instance
        role: offline
        rules:
          - host    all     +dbrole_offline               10.0.0.0/8        md5
          - host    all     +dbrole_offline               172.16.0.0/12     md5
          - host    all     +dbrole_offline               192.168.0.0/16    md5

    pg_hba_rules_extra: []                        # extra hba rules (for cluster/instance overwrite)

    pgbouncer_hba_rules:                          # pgbouncer host-based authentication rules
      - title: local password access
        role: common
        rules:
          - local  all          all                                     md5
          - host   all          all                     127.0.0.1/32    md5

      - title: intranet password access
        role: common
        rules:
          - host   all          all                     10.0.0.0/8      md5
          - host   all          all                     172.16.0.0/12   md5
          - host   all          all                     192.168.0.0/16  md5

    pgbouncer_hba_rules_extra: []                 # extra pgbouncer hba rules (for cluster/instance overwrite)
    # pg_users: []                                # business users
    # pg_databases: []                            # business databases

    #------------------------------------------------------------------------------
    # MONITOR PROVISION
    #------------------------------------------------------------------------------
    # - install - #
    exporter_install: none                        # none|yum|binary, none by default
    exporter_repo_url: ''                         # if set, repo will be added to /etc/yum.repos.d/ before yum installation

    # - collect - #
    exporter_metrics_path: /metrics               # default metric path for pg related exporter

    # - node exporter - #
    node_exporter_enabled: true                   # setup node_exporter on instance
    node_exporter_port: 9100                      # default port for node exporter
    node_exporter_options: '--no-collector.softnet --collector.systemd --collector.ntp --collector.tcpstat --collector.processes'

    # - pg exporter - #
    pg_exporter_config: pg_exporter-demo.yaml     # default config files for pg_exporter
    pg_exporter_enabled: true                     # setup pg_exporter on instance
    pg_exporter_port: 9630                        # default port for pg exporter
    pg_exporter_url: ''                           # optional, if not set, generate from reference parameters

    # - pgbouncer exporter - #
    pgbouncer_exporter_enabled: true              # setup pgbouncer_exporter on instance (if you don't have pgbouncer, disable it)
    pgbouncer_exporter_port: 9631                 # default port for pgbouncer exporter
    pgbouncer_exporter_url: ''                    # optional, if not set, generate from reference parameters

    # - promtail - #                              # promtail is a beta feature which requires manual deployment
    promtail_enabled: true                        # enable promtail logging collector?
    promtail_clean: false                         # remove promtail status file? false by default
    promtail_port: 9080                           # default listen address for promtail
    promtail_status_file: /tmp/promtail-status.yml
    promtail_send_url: http://10.10.10.10:3100/loki/api/v1/push  # loki url to receive logs

    #------------------------------------------------------------------------------
    # SERVICE PROVISION
    #------------------------------------------------------------------------------
    pg_weight: 100              # default load balance weight (instance level)

    # - service - #
    pg_services:                                  # how to expose postgres service in cluster?
      # primary service will route {ip|name}:5433 to primary pgbouncer (5433->6432 rw)
      - name: primary           # service name {{ pg_cluster }}-primary
        src_ip: "*"
        src_port: 5433
        dst_port: pgbouncer     # 5433 route to pgbouncer
        check_url: /primary     # primary health check, success when instance is primary
        selector: "[]"          # select all instance as primary service candidate

      # replica service will route {ip|name}:5434 to replica pgbouncer (5434->6432 ro)
      - name: replica           # service name {{ pg_cluster }}-replica
        src_ip: "*"
        src_port: 5434
        dst_port: pgbouncer
        check_url: /read-only   # read-only health check. (including primary)
        selector: "[]"          # select all instance as replica service candidate
        selector_backup: "[? pg_role == `primary`]"   # primary are used as backup server in replica service

      # default service will route {ip|name}:5436 to primary postgres (5436->5432 primary)
      - name: default           # service's actual name is {{ pg_cluster }}-default
        src_ip: "*"             # service bind ip address, * for all, vip for cluster virtual ip address
        src_port: 5436          # bind port, mandatory
        dst_port: postgres      # target port: postgres|pgbouncer|port_number , pgbouncer(6432) by default
        check_method: http      # health check method: only http is available for now
        check_port: patroni     # health check port:  patroni|pg_exporter|port_number , patroni by default
        check_url: /primary     # health check url path, / as default
        check_code: 200         # health check http code, 200 as default
        selector: "[]"          # instance selector
        haproxy:                # haproxy specific fields
          maxconn: 3000         # default front-end connection
          balance: roundrobin   # load balance algorithm (roundrobin by default)
          default_server_options: 'inter 3s fastinter 1s downinter 5s rise 3 fall 3 on-marked-down shutdown-sessions slowstart 30s maxconn 3000 maxqueue 128 weight 100'

      # offline service will route {ip|name}:5438 to offline postgres (5438->5432 offline)
      - name: offline           # service name {{ pg_cluster }}-offline
        src_ip: "*"
        src_port: 5438
        dst_port: postgres
        check_url: /replica     # offline MUST be a replica
        selector: "[? pg_role == `offline` || pg_offline_query ]"         # instances with pg_role == 'offline' or instance marked with 'pg_offline_query == true'
        selector_backup: "[? pg_role == `replica` && !pg_offline_query]"  # replica are used as backup server in offline service

    pg_services_extra: []        # extra services to be added

    # - haproxy - #
    haproxy_enabled: true                         # enable haproxy among every cluster members
    haproxy_reload: true                          # reload haproxy after config
    haproxy_admin_auth_enabled: false             # enable authentication for haproxy admin?
    haproxy_admin_username: admin                 # default haproxy admin username
    haproxy_admin_password: admin                 # default haproxy admin password
    haproxy_exporter_port: 9101                   # default admin/exporter port
    haproxy_client_timeout: 3h                    # client side connection timeout
    haproxy_server_timeout: 3h                    # server side connection timeout

    # - vip - #
    vip_mode: none                                # none | l2 | l4
    vip_reload: true                              # whether reload service after config
    # vip_address: 127.0.0.1                      # virtual ip address ip (l2 or l4)
    # vip_cidrmask: 24                            # virtual ip address cidr mask (l2 only)
    # vip_interface: eth0                         # virtual ip network interface (l2 only)

    # - dns - #                                   # NOT IMPLEMENTED
    # dns_mode: vip                               # vip|all|selector: how to resolve cluster DNS?
    # dns_selector: '[]'                          # if dns_mode == vip, filter instances been resolved

...

tuned-adm profile oltp    # 启用OLTP模式
tuned-adm profile olap    # 启用OLAP模式
tuned-adm profile crit    # 启用CRIT模式
tuned-adm profile tiny    # 启用TINY模式

# 如需启动 tuned，请以 root 身份运行下列指令：
systemctl start tuned

# 若要在每次计算机启动时激活 tuned，请输入以下指令：
systemctl enable tuned

# 其它的 tuned 控制，例如配置文件选择等，请使用：
tuned-adm

# 若要查看可用的已安装配置文件，此命令需要 tuned 服务正在运行。
tuned-adm list

# 若要查看目前已激活的配置文件，请运行：
tuned-adm active

# 若要选择或激活某一配置文件，请运行：
tuned-adm profile profile
# 例如
tuned-adm profile powersave

# 若要让 tuned 推荐最适合您的系统的配置文件，同时不改变任何现有的配置文件，也不使用安装期间使用过的逻辑，请运行以下指令：
tuned-adm recommend

# 要禁用所有微调：
tuned-adm off

要列出所有可用配置文件并识别目前激活的配置文件，请运行：
tuned-adm list
要只显示当前激活的配置文件请运行：
tuned-adm active
要切换到某个可用的配置文件请运行：
tuned-adm profile profile_name
例如：
tuned-adm profile server-powersave

# tuned configuration
#==============================================================#
# File      :   tuned.conf
# Mtime     :   2020-06-29
# Desc      :   Tune operatiing system to oltp mode
# Path      :   /etc/tuned/oltp/tuned.conf
# Author    :   Vonng(fengruohang@outlook.com)
# Copyright (C) 2019-2020 Ruohang Feng
#==============================================================#

[main]
summary=Optimize for PostgreSQL OLTP System
include=network-latency

[cpu]
force_latency=1
governor=performance
energy_perf_bias=performance
min_perf_pct=100

[vm]
# disable transparent hugepages
transparent_hugepages=never

[sysctl]
#-------------------------------------------------------------#
#                           KERNEL                            #
#-------------------------------------------------------------#
# disable numa balancing
kernel.numa_balancing=0

# total shmem size in bytes： $(expr $(getconf _PHYS_PAGES) / 2 \* $(getconf PAGE_SIZE))
{% if param_shmall is defined and param_shmall != '' %}
kernel.shmall = {{ param_shmall }}
{% endif %}

# total shmem size in pages:  $(expr $(getconf _PHYS_PAGES) / 2)
{% if param_shmmax is defined and param_shmmax != '' %}
kernel.shmmax = {{ param_shmmax }}
{% endif %}

# total shmem segs 4096 -> 8192
kernel.shmmni=8192

# total msg queue number, set to mem size in MB
kernel.msgmni=32768

# max length of message queue
kernel.msgmnb=65536

# max size of message
kernel.msgmax=65536

kernel.pid_max=131072

# max(Sem in Set)=2048, max(Sem)=max(Sem in Set) x max(SemSet) , max(Sem per Ops)=2048, max(SemSet)=65536
kernel.sem=2048 134217728 2048 65536

# do not sched postgres process in group
kernel.sched_autogroup_enabled = 0

# total time the scheduler will consider a migrated process cache hot and, thus, less likely to be remigrated
# defaut = 0.5ms (500000ns), update to 5ms , depending on your typical query (e.g < 1ms)
kernel.sched_migration_cost_ns=5000000

#-------------------------------------------------------------#
#                             VM                              #
#-------------------------------------------------------------#
# try not using swap
vm.swappiness=0

# disable when most mem are for file cache
vm.zone_reclaim_mode=0

# overcommit threshhold = 80%
vm.overcommit_memory=2
vm.overcommit_ratio=80

# vm.dirty_background_bytes=67108864 # 64MB mem (2xRAID cache) wake the bgwriter
vm.dirty_background_ratio=3       # latency-performance default
vm.dirty_ratio=10                 # latency-performance default

# deny access on 0x00000 - 0x10000
vm.mmap_min_addr=65536

#-------------------------------------------------------------#
#                        Filesystem                           #
#-------------------------------------------------------------#
# max open files: 382589 -> 167772160
fs.file-max=167772160

# max concurrent unfinished async io, should be larger than 1M.  65536->1M
fs.aio-max-nr=1048576


#-------------------------------------------------------------#
#                          Network                            #
#-------------------------------------------------------------#
# max connection in listen queue (triggers retrans if full)
net.core.somaxconn=65535
net.core.netdev_max_backlog=8192
# tcp receive/transmit buffer default = 256KiB
net.core.rmem_default=262144
net.core.wmem_default=262144
# receive/transmit buffer limit = 4MiB
net.core.rmem_max=4194304
net.core.wmem_max=4194304

# ip options
net.ipv4.ip_forward=1
net.ipv4.ip_nonlocal_bind=1
net.ipv4.ip_local_port_range=32768 65000

# tcp options
net.ipv4.tcp_timestamps=1
net.ipv4.tcp_tw_reuse=1
net.ipv4.tcp_tw_recycle=0
net.ipv4.tcp_syncookies=0
net.ipv4.tcp_synack_retries=1
net.ipv4.tcp_syn_retries=1

# tcp read/write buffer
net.ipv4.tcp_rmem="4096 87380 16777216"
net.ipv4.tcp_wmem="4096 16384 16777216"
net.ipv4.udp_mem="3145728 4194304 16777216"

# tcp probe fail interval: 75s -> 20s
net.ipv4.tcp_keepalive_intvl=20
# tcp break after 3 * 20s = 1m
net.ipv4.tcp_keepalive_probes=3
# probe peroid = 1 min
net.ipv4.tcp_keepalive_time=60

net.ipv4.tcp_fin_timeout=5
net.ipv4.tcp_max_tw_buckets=262144
net.ipv4.tcp_max_syn_backlog=8192
net.ipv4.neigh.default.gc_thresh1=80000
net.ipv4.neigh.default.gc_thresh2=90000
net.ipv4.neigh.default.gc_thresh3=100000

net.bridge.bridge-nf-call-iptables=1
net.bridge.bridge-nf-call-ip6tables=1
net.bridge.bridge-nf-call-arptables=1

# max connection tracking number
net.netfilter.nf_conntrack_max=1048576

# tuned configuration
#==============================================================#
# File      :   tuned.conf
# Mtime     :   2020-09-18
# Desc      :   Tune operatiing system to olap mode
# Path      :   /etc/tuned/olap/tuned.conf
# Author    :   Vonng(fengruohang@outlook.com)
# Copyright (C) 2019-2020 Ruohang Feng
#==============================================================#

[main]
summary=Optimize for PostgreSQL OLAP System
include=network-throughput

[cpu]
force_latency=1
governor=performance
energy_perf_bias=performance
min_perf_pct=100

[vm]
# disable transparent hugepages
transparent_hugepages=never

[sysctl]
#-------------------------------------------------------------#
#                           KERNEL                            #
#-------------------------------------------------------------#
# disable numa balancing
kernel.numa_balancing=0

# total shmem size in bytes： $(expr $(getconf _PHYS_PAGES) / 2 \* $(getconf PAGE_SIZE))
{% if param_shmall is defined and param_shmall != '' %}
kernel.shmall = {{ param_shmall }}
{% endif %}

# total shmem size in pages:  $(expr $(getconf _PHYS_PAGES) / 2)
{% if param_shmmax is defined and param_shmmax != '' %}
kernel.shmmax = {{ param_shmmax }}
{% endif %}

# total shmem segs 4096 -> 8192
kernel.shmmni=8192

# total msg queue number, set to mem size in MB
kernel.msgmni=32768

# max length of message queue
kernel.msgmnb=65536

# max size of message
kernel.msgmax=65536

kernel.pid_max=131072

# max(Sem in Set)=2048, max(Sem)=max(Sem in Set) x max(SemSet) , max(Sem per Ops)=2048, max(SemSet)=65536
kernel.sem=2048 134217728 2048 65536

# do not sched postgres process in group
kernel.sched_autogroup_enabled = 0

# total time the scheduler will consider a migrated process cache hot and, thus, less likely to be remigrated
# defaut = 0.5ms (500000ns), update to 5ms , depending on your typical query (e.g < 1ms)
kernel.sched_migration_cost_ns=5000000

#-------------------------------------------------------------#
#                             VM                              #
#-------------------------------------------------------------#
# try not using swap
# vm.swappiness=10

# disable when most mem are for file cache
vm.zone_reclaim_mode=0

# overcommit threshhold = 80%
vm.overcommit_memory=2
vm.overcommit_ratio=80

vm.dirty_background_ratio = 10    # throughput-performance default
vm.dirty_ratio=80                 # throughput-performance default 40 -> 80

# deny access on 0x00000 - 0x10000
vm.mmap_min_addr=65536

#-------------------------------------------------------------#
#                        Filesystem                           #
#-------------------------------------------------------------#
# max open files: 382589 -> 167772160
fs.file-max=167772160

# max concurrent unfinished async io, should be larger than 1M.  65536->1M
fs.aio-max-nr=1048576


#-------------------------------------------------------------#
#                          Network                            #
#-------------------------------------------------------------#
# max connection in listen queue (triggers retrans if full)
net.core.somaxconn=65535
net.core.netdev_max_backlog=8192
# tcp receive/transmit buffer default = 256KiB
net.core.rmem_default=262144
net.core.wmem_default=262144
# receive/transmit buffer limit = 4MiB
net.core.rmem_max=4194304
net.core.wmem_max=4194304

# ip options
net.ipv4.ip_forward=1
net.ipv4.ip_nonlocal_bind=1
net.ipv4.ip_local_port_range=32768 65000

# tcp options
net.ipv4.tcp_timestamps=1
net.ipv4.tcp_tw_reuse=1
net.ipv4.tcp_tw_recycle=0
net.ipv4.tcp_syncookies=0
net.ipv4.tcp_synack_retries=1
net.ipv4.tcp_syn_retries=1

# tcp read/write buffer
net.ipv4.tcp_rmem="4096 87380 16777216"
net.ipv4.tcp_wmem="4096 16384 16777216"
net.ipv4.udp_mem="3145728 4194304 16777216"

# tcp probe fail interval: 75s -> 20s
net.ipv4.tcp_keepalive_intvl=20
# tcp break after 3 * 20s = 1m
net.ipv4.tcp_keepalive_probes=3
# probe peroid = 1 min
net.ipv4.tcp_keepalive_time=60

net.ipv4.tcp_fin_timeout=5
net.ipv4.tcp_max_tw_buckets=262144
net.ipv4.tcp_max_syn_backlog=8192
net.ipv4.neigh.default.gc_thresh1=80000
net.ipv4.neigh.default.gc_thresh2=90000
net.ipv4.neigh.default.gc_thresh3=100000

net.bridge.bridge-nf-call-iptables=1
net.bridge.bridge-nf-call-ip6tables=1
net.bridge.bridge-nf-call-arptables=1

# max connection tracking number
net.netfilter.nf_conntrack_max=1048576

# tuned configuration
#==============================================================#
# File      :   tuned.conf
# Mtime     :   2020-06-29
# Desc      :   Tune operatiing system to crit mode
# Path      :   /etc/tuned/crit/tuned.conf
# Author    :   Vonng(fengruohang@outlook.com)
# Copyright (C) 2019-2020 Ruohang Feng
#==============================================================#

[main]
summary=Optimize for PostgreSQL CRIT System
include=network-latency

[cpu]
force_latency=1
governor=performance
energy_perf_bias=performance
min_perf_pct=100

[vm]
# disable transparent hugepages
transparent_hugepages=never

[sysctl]
#-------------------------------------------------------------#
#                           KERNEL                            #
#-------------------------------------------------------------#
# disable numa balancing
kernel.numa_balancing=0

# total shmem size in bytes： $(expr $(getconf _PHYS_PAGES) / 2 \* $(getconf PAGE_SIZE))
{% if param_shmall is defined and param_shmall != '' %}
kernel.shmall = {{ param_shmall }}
{% endif %}

# total shmem size in pages:  $(expr $(getconf _PHYS_PAGES) / 2)
{% if param_shmmax is defined and param_shmmax != '' %}
kernel.shmmax = {{ param_shmmax }}
{% endif %}

# total shmem segs 4096 -> 8192
kernel.shmmni=8192

# total msg queue number, set to mem size in MB
kernel.msgmni=32768

# max length of message queue
kernel.msgmnb=65536

# max size of message
kernel.msgmax=65536

kernel.pid_max=131072

# max(Sem in Set)=2048, max(Sem)=max(Sem in Set) x max(SemSet) , max(Sem per Ops)=2048, max(SemSet)=65536
kernel.sem=2048 134217728 2048 65536

# do not sched postgres process in group
kernel.sched_autogroup_enabled = 0

# total time the scheduler will consider a migrated process cache hot and, thus, less likely to be remigrated
# defaut = 0.5ms (500000ns), update to 5ms , depending on your typical query (e.g < 1ms)
kernel.sched_migration_cost_ns=5000000

#-------------------------------------------------------------#
#                             VM                              #
#-------------------------------------------------------------#
# try not using swap
vm.swappiness=0

# disable when most mem are for file cache
vm.zone_reclaim_mode=0

# overcommit threshhold = 80%
vm.overcommit_memory=2
vm.overcommit_ratio=100

# 64MB mem (2xRAID cache) wake the bgwriter
vm.dirty_background_bytes=67108864
# vm.dirty_background_ratio=3       # latency-performance default
vm.dirty_ratio=6                    # latency-performance default

# deny access on 0x00000 - 0x10000
vm.mmap_min_addr=65536

#-------------------------------------------------------------#
#                        Filesystem                           #
#-------------------------------------------------------------#
# max open files: 382589 -> 167772160
fs.file-max=167772160

# max concurrent unfinished async io, should be larger than 1M.  65536->1M
fs.aio-max-nr=1048576


#-------------------------------------------------------------#
#                          Network                            #
#-------------------------------------------------------------#
# max connection in listen queue (triggers retrans if full)
net.core.somaxconn=65535
net.core.netdev_max_backlog=8192
# tcp receive/transmit buffer default = 256KiB
net.core.rmem_default=262144
net.core.wmem_default=262144
# receive/transmit buffer limit = 4MiB
net.core.rmem_max=4194304
net.core.wmem_max=4194304

# ip options
net.ipv4.ip_forward=1
net.ipv4.ip_nonlocal_bind=1
net.ipv4.ip_local_port_range=32768 65000

# tcp options
net.ipv4.tcp_timestamps=1
net.ipv4.tcp_tw_reuse=1
net.ipv4.tcp_tw_recycle=0
net.ipv4.tcp_syncookies=0
net.ipv4.tcp_synack_retries=1
net.ipv4.tcp_syn_retries=1

# tcp read/write buffer
net.ipv4.tcp_rmem="4096 87380 16777216"
net.ipv4.tcp_wmem="4096 16384 16777216"
net.ipv4.udp_mem="3145728 4194304 16777216"

# tcp probe fail interval: 75s -> 20s
net.ipv4.tcp_keepalive_intvl=20
# tcp break after 3 * 20s = 1m
net.ipv4.tcp_keepalive_probes=3
# probe peroid = 1 min
net.ipv4.tcp_keepalive_time=60

net.ipv4.tcp_fin_timeout=5
net.ipv4.tcp_max_tw_buckets=262144
net.ipv4.tcp_max_syn_backlog=8192
net.ipv4.neigh.default.gc_thresh1=80000
net.ipv4.neigh.default.gc_thresh2=90000
net.ipv4.neigh.default.gc_thresh3=100000

net.bridge.bridge-nf-call-iptables=1
net.bridge.bridge-nf-call-ip6tables=1
net.bridge.bridge-nf-call-arptables=1

# max connection tracking number
net.netfilter.nf_conntrack_max=1048576

# tuned configuration
#==============================================================#
# File      :   tuned.conf
# Mtime     :   2020-06-29
# Desc      :   Tune operatiing system to tiny mode
# Path      :   /etc/tuned/tiny/tuned.conf
# Author    :   Vonng(fengruohang@outlook.com)
# Copyright (C) 2019-2020 Ruohang Feng
#==============================================================#

[main]
summary=Optimize for PostgreSQL TINY System
# include=virtual-guest

[vm]
# disable transparent hugepages
transparent_hugepages=never

[sysctl]
#-------------------------------------------------------------#
#                           KERNEL                            #
#-------------------------------------------------------------#
# disable numa balancing
kernel.numa_balancing=0

# If a workload mostly uses anonymous memory and it hits this limit, the entire
# working set is buffered for I/O, and any more write buffering would require
# swapping, so it's time to throttle writes until I/O can catch up.  Workloads
# that mostly use file mappings may be able to use even higher values.
#
# The generator of dirty data starts writeback at this percentage (system default
# is 20%)
vm.dirty_ratio = 40

# Filesystem I/O is usually much more efficient than swapping, so try to keep
# swapping low.  It's usually safe to go even lower than this on systems with
# server-grade storage.
vm.swappiness = 30

#-------------------------------------------------------------#
#                          Network                            #
#-------------------------------------------------------------#
# tcp options
net.ipv4.tcp_timestamps=1
net.ipv4.tcp_tw_reuse=1
net.ipv4.tcp_tw_recycle=0
net.ipv4.tcp_syncookies=0
net.ipv4.tcp_synack_retries=1
net.ipv4.tcp_syn_retries=1

# tcp probe fail interval: 75s -> 20s
net.ipv4.tcp_keepalive_intvl=20
# tcp break after 3 * 20s = 1m
net.ipv4.tcp_keepalive_probes=3
# probe peroid = 1 min
net.ipv4.tcp_keepalive_time=60

# Database kernel optimisation
fs.aio-max-nr = 1048576 # 限制并发未完成的异步请求数目，，不应小于1M
fs.file-max = 16777216  # 最大打开16M个文件

# kernel
kernel.shmmax = 485058		# 共享内存最大页面数量: $(expr $(getconf _PHYS_PAGES) / 2)
kernel.shmall = 1986797568 	# 共享内存总大小： $(expr $(getconf _PHYS_PAGES) / 2 \* $(getconf PAGE_SIZE))
kernel.shmmni = 16384 		# 系统范围内共享内存段的最大数量 4096 -> 16384
kernel.msgmni = 32768		# 系统的消息队列数目,影响可以启动的代理程序数 设为内存MB数
kernel.msgmnb = 65536		# 影响队列的大小
kernel.msgmax = 65536		# 影响队列中可以发送的消息的大小
kernel.numa_balancing = 0   # Numa禁用
kernel.sched_migration_cost_ns = 5000000 # 5ms内，调度认为进程还是Hot的。
kernel.sem = 2048 134217728 2048 65536   # 每个信号集最大信号量2048，系统总共可用信号量134217728，单次最大操作2048，信号集总数65536

# vm
vm.dirty_ratio = 80                       # 绝对限制，超过80%阻塞写请求刷盘
vm.dirty_background_bytes = 268435456     # 256MB脏数据唤醒刷盘进程
vm.dirty_expire_centisecs = 6000          # 1分钟前的数据被认为需要刷盘
vm.dirty_writeback_centisecs= 500         # 刷新进程运行间隔5秒
vm.mmap_min_addr = 65536                  # 禁止访问0x10000下的内存
vm.zone_reclaim_mode = 0                  # Numa禁用

# vm swap
vm.swappiness = 0                         # 禁用SWAP，但高水位仍会有
vm.overcommit_memory = 2                  # 允许一定程度的Overcommit
vm.overcommit_ratio = 50                  # 允许的Overcommit:$((($mem - $swap) * 100 / $mem))

# tcp memory
net.ipv4.tcp_rmem = 8192 65536 16777216		# tcp读buffer: 32M/256M/16G
net.ipv4.tcp_wmem = 8192 65536 16777216		# tcp写buffer: 32M/256M/16G
net.ipv4.tcp_mem = 131072 262144 16777216	# tcp 内存使用 512M/1G/16G
net.core.rmem_default = 262144      		# 接受缓冲区默认大小: 256K
net.core.rmem_max = 4194304         		# 接受缓冲区最大大小: 4M
net.core.wmem_default = 262144      		# 发送缓冲区默认大小: 256K
net.core.wmem_max = 4194304         		# 发送缓冲区最大大小: 4M
# tcp keepalive
net.ipv4.tcp_keepalive_intvl = 20	# 探测没有确认时，重新发送探测的频度。默认75s -> 20s
net.ipv4.tcp_keepalive_probes = 3	# 3 * 20 = 1分钟超时断开
net.ipv4.tcp_keepalive_time = 60	# 探活周期1分钟
# tcp port resure
net.ipv4.tcp_tw_reuse = 1           # 允许将TIME_WAIT socket用于新的TCP连接。默认为0
net.ipv4.tcp_tw_recycle = 0			# 快速回收，已弃用
net.ipv4.tcp_fin_timeout = 5        # 保持在FIN-WAIT-2状态的秒时间
net.ipv4.tcp_timestamps = 1
# tcp anti-flood
net.ipv4.tcp_syncookies = 1			# SYN_RECV队列满后发cookie，防止恶意攻击
net.ipv4.tcp_synack_retries = 1		# 收到不完整sync后的重试次数 5->2
net.ipv4.tcp_syn_retries = 1         #表示在内核放弃建立连接之前发送SYN包的数量。
# tcp load-balancer
net.ipv4.ip_forward = 1						# IP转发
net.ipv4.ip_nonlocal_bind = 1				# 绑定非本机地址
net.netfilter.nf_conntrack_max = 1048576	# 最大跟踪连接数
net.ipv4.ip_local_port_range = 10000 65535	# 端口范围
net.ipv4.tcp_max_tw_buckets = 262144		# 256k  TIME_WAIT
net.core.somaxconn = 65535          		# 限制LISTEN队列最大数据包量，触发重传机制。
net.ipv4.tcp_max_syn_backlog = 8192 		# SYN队列大小：1024->8192
net.core.netdev_max_backlog = 8192			# 网卡收包快于内核时，允许队列长度

---
  - name: node-rules
    rules:
      #==============================================================#
      #                         Aliveness                            #
      #==============================================================#
      # TODO: change this to your node exporter port
      - record: node_exporter_up
        expr: up{instance=~".*:9099"}
      - record: node:uptime
        expr: time() - node_boot_time_seconds{}


      #==============================================================#
      #                             CPU                              #
      #==============================================================#
      # cpu mode time ratio
      - record: node:cpu:cpu_mode
        expr: irate(node_cpu_seconds_total{}[1m])
      - record: node:ins:cpu_mode
        expr: sum without (cpu) (node:cpu:cpu_mode)
      - record: node:cls:cpu_mode
        expr: sum by (cls, mode) (node:ins:cpu_mode)

      # cpu schedule time-slices
      - record: node:cpu:sched_timeslices
        expr: irate(node_schedstat_timeslices_total{}[1m])
      - record: node:ins:sched_timeslices
        expr: sum without (cpu) (node:cpu:sched_timeslices)
      - record: node:cls:sched_timeslicesa
        expr: sum by (cls) (node:ins:sched_timeslices)

      # cpu count
      - record: node:ins:cpu_count
        expr: count without (cpu) (node:cpu:cpu_usage)
      - record: node:cls:cpu_count
        expr: sum by (cls) (node:ins:cpu_count)

      # cpu usage
      - record: node:cpu:cpu_usage
        expr: 1 - sum without (mode) (node:cpu:cpu_mode{mode="idle"})
      - record: node:ins:cpu_usage
        expr: sum without (cpu) (node:cpu:cpu_usage) / node:ins:cpu_count
      - record: node:cls:cpu_usage
        expr: sum by (cls) (node:ins:cpu_usage * node:ins:cpu_count) / sum by (cls) (node:ins:cpu_count)

      # cpu usage avg5m
      - record: node:cpu:cpu_usage_avg5m
        expr: avg_over_time(node:cpu:cpu_usage[5m])
      - record: node:ins:cpu_usage_avg5m
        expr: avg_over_time(node:ins:cpu_usage[5m])
      - record: node:cls:cpu_usage_avg5m
        expr: avg_over_time(node:cls:cpu_usage[5m])

      #==============================================================#
      #                            Memory                            #
      #==============================================================#
      # mem usage
      - record: node:ins:mem_app
        expr: node_memory_MemTotal_bytes - node_memory_MemFree_bytes - node_memory_Buffers_bytes - node_memory_Cached_bytes - node_memory_Slab_bytes - node_memory_PageTables_bytes - node_memory_SwapCached_bytes
      - record: node:ins:mem_free
        expr: node_memory_MemFree_bytes{} + node_memory_Cached_bytes{}
      - record: node:ins:mem_usage
        expr: node:ins:mem_app / node_memory_MemTotal_bytes
      - record: node:cls:mem_usage
        expr: sum by (cls) (node:ins:mem_app) / sum by (cls) (node_memory_MemTotal_bytes)
      - record: node:ins:swap_usage
        expr: 1 - node_memory_SwapFree_bytes{} / node_memory_SwapTotal_bytes{}


      #==============================================================#
      #                            Disk                              #
      #==============================================================#
      # disk read iops
      - record: node:dev:disk_read_iops
        expr: irate(node_disk_reads_completed_total{device=~"[a-zA-Z-_]+"}[1m])
      - record: node:ins:disk_read_iops
        expr: sum without (device) (node:dev:disk_read_iops)
      - record: node:cls:disk_read_iops
        expr: sum by (cls) (node:ins:disk_read_iops)

      # disk write iops
      - record: node:dev:disk_write_iops
        expr: irate(node_disk_writes_completed_total{device=~"[a-zA-Z-_]+"}[1m])
      - record: node:ins:disk_write_iops
        expr: sum without (device) (node:dev:disk_write_iops)
      - record: node:cls:disk_write_iops
        expr: sum by (cls) (node:ins:disk_write_iops)

      # disk iops
      - record: node:dev:disk_iops
        expr: node:dev:disk_read_iops + node:dev:disk_write_iops
      - record: node:ins:disk_iops
        expr: node:ins:disk_read_iops + node:ins:disk_write_iops
      - record: node:cls:disk_iops
        expr: node:cls:disk_read_iops + node:cls:disk_write_iops

      # read bandwidth (rate1m)
      - record: node:dev:disk_read_rate
        expr: rate(node_disk_read_bytes_total{device=~"[a-zA-Z-_]+"}[1m])
      - record: node:ins:disk_read_rate
        expr: sum without (device) (node:dev:disk_read_rate)
      - record: node:cls:disk_read_rate
        expr: sum by (cls) (node:ins:disk_read_rate)

      # write bandwidth (rate1m)
      - record: node:dev:disk_write_rate
        expr: rate(node_disk_written_bytes_total{device=~"[a-zA-Z-_]+"}[1m])
      - record: node:ins:disk_write_rate
        expr: sum without (device) (node:dev:disk_write_rate)
      - record: node:cls:disk_write_rate
        expr: sum by (cls) (node:ins:disk_write_rate)

      # io bandwidth (rate1m)
      - record: node:dev:disk_io_rate
        expr: node:dev:disk_read_rate + node:dev:disk_write_rate
      - record: node:ins:disk_io_rate
        expr: node:ins:disk_read_rate + node:ins:disk_write_rate
      - record: node:cls:disk_io_rate
        expr: node:cls:disk_read_rate + node:cls:disk_write_rate

      # read/write total time
      - record: node:dev:disk_read_time
        expr: rate(node_disk_read_time_seconds_total{device=~"[a-zA-Z-_]+"}[1m])
      - record: node:dev:disk_write_time
        expr: rate(node_disk_read_time_seconds_total{device=~"[a-zA-Z-_]+"}[1m])

      # read/write response time
      - record: node:dev:disk_read_rt
        expr: node:dev:disk_read_time / node:dev:disk_read_iops
      - record: node:dev:disk_write_rt
        expr: node:dev:disk_write_time / node:dev:disk_write_iops
      - record: node:dev:disk_rt
        expr: (node:dev:disk_read_time + node:dev:disk_write_time) / node:dev:iops


      #==============================================================#
      #                            Network                           #
      #==============================================================#
      # transmit bandwidth (out)
      - record: node:dev:network_tx
        expr: irate(node_network_transmit_bytes_total{}[1m])
      - record: node:ins:network_tx
        expr: sum without (device) (node:dev:network_tx{device!~"lo|bond.*"})
      - record: node:cls:network_tx
        expr: sum by (cls) (node:ins:network_tx)

      # receive bandwidth (in)
      - record: node:dev:network_rx
        expr: irate(node_network_receive_bytes_total{}[1m])
      - record: node:ins:network_rx
        expr: sum without (device) (node:dev:network_rx{device!~"lo|bond.*"})
      - record: node:cls:network_rx
        expr: sum by (cls) (node:ins:network_rx)

      # io bandwidth
      - record: node:dev:network_io_rate
        expr: node:dev:network_tx + node:dev:network_rx
      - record: node:ins:network_io
        expr: node:ins:network_tx + node:ins:network_rx
      - record: node:cls:network_io
        expr: node:cls:network_tx + node:cls:network_rx


      #==============================================================#
      #                           Schedule                           #
      #==============================================================#
      # normalized load
      - record: node:ins:stdload1
        expr: node_load1 / node:ins:cpu_count
      - record: node:ins:stdload5
        expr: node_load5 / node:ins:cpu_count
      - record: node:ins:stdload15
        expr: node_load15 / node:ins:cpu_count

      # process
      - record: node:ins:forks
        expr: irate(node_forks_total[1m])
      # interrupt & context switch
      - record: node:ins:intrrupt
        expr: irate(node_intr_total[1m])
      - record: node:ins:ctx_switch
        expr: irate(node_context_switches_total{}[1m])


      #==============================================================#
      #                              VM                              #
      #==============================================================#
      - record: node:ins:pagefault
        expr: irate(node_vmstat_pgfault[1m])
      - record: node:ins:pagein
        expr: irate(node_vmstat_pgpgin[1m])
      - record: node:ins:pageout
        expr: irate(node_vmstat_pgpgout[1m])
      - record: node:ins:swapin
        expr: irate(node_vmstat_pswpin[1m])
      - record: node:ins:swapout
        expr: irate(node_vmstat_pswpout[1m])


      #==============================================================#
      #                              FS                              #
      #==============================================================#
      # filesystem space usage
      - record: node:fs:free_bytes
        expr: max without(device, fstype) (node_filesystem_free_bytes{fstype!~"(n|root|tmp)fs.*"})
      - record: node:fs:avail_bytes
        expr: max without(device, fstype) (node_filesystem_avail_bytes{fstype!~"(n|root|tmp)fs.*"})
      - record: node:fs:size_bytes
        expr: max without(device, fstype) (node_filesystem_size_bytes{fstype!~"(n|root|tmp)fs.*"})
      - record: node:fs:space_usage
        expr: 1 - (node:fs:avail_bytes{} / node:fs:size_bytes{})
      - record: node:fs:free_inode
        expr: max without(device, fstype) (node_filesystem_files_free{fstype!~"(n|root|tmp)fs.*"})
      - record: node:fs:total_inode
        expr: max without(device, fstype) (node_filesystem_files{fstype!~"(n|root|tmp)fs.*"})

      # space delta and prediction
      - record: node:fs:space_deriv_1h
        expr: 0 - deriv(node_filesystem_avail_bytes{}[1h])
      - record: node:fs:space_exhaust
        expr: (node_filesystem_avail_bytes{} / node:fs:space_deriv_1h{}) > 0

      # fs inode usage
      - record: node:fs:inode_usage
        expr: 1 - (node:fs:free_inode / node:fs:total_inode)
      # file descriptor usage
      - record: node:ins:fd_usage
        expr: node_filefd_allocated / node_filefd_maximum


      #==============================================================#
      #                             TCP                              #
      #==============================================================#
      # tcp segments (rate1m)
      - record: node:ins:tcp_insegs
        expr: rate(node_netstat_Tcp_InSegs{}[1m])
      - record: node:ins:tcp_outsegs
        expr: rate(node_netstat_Tcp_OutSegs{}[1m])
      - record: node:ins:tcp_retranssegs
        expr: rate(node_netstat_Tcp_RetransSegs{}[1m])
      - record: node:ins:tcp_segs
        expr: node:ins:tcp_insegs + node:ins:tcp_outsegs
      # retransmit
      - record: node:ins:tcp_retrans_rate
        expr: node:ins:tcp_retranssegs / node:ins:tcp_outsegs
      # overflow
      - record: node:ins:tcp_overflow_rate
        expr: rate(node_netstat_TcpExt_ListenOverflows[1m])


      #==============================================================#
      #                           Netstat                            #
      #==============================================================#
      # tcp open (rate1m)
      - record: node:ins:tcp_passive_opens
        expr: rate(node_netstat_Tcp_PassiveOpens[1m])
      - record: node:ins:tcp_active_opens
        expr: rate(node_netstat_Tcp_ActiveOpens[1m])
      # tcp close
      - record: node:ins:tcp_attempt_fails
        expr: rate(node_netstat_Tcp_AttemptFails[1m])
      - record: node:ins:tcp_estab_resets
        expr: rate(node_netstat_Tcp_EstabResets[1m])
      # tcp drop
      - record: node:ins:tcp_overflow
        expr: rate(node_netstat_TcpExt_ListenOverflows[1m])
      - record: node:ins:tcp_dropped
        expr: rate(node_netstat_TcpExt_ListenDrops[1m])


      #==============================================================#
      #                             NTP                              #
      #==============================================================#
      - record: node:cls:ntp_offset_range
        expr: max by (cls)(node_ntp_offset_seconds) - min by (cls)(node_ntp_offset_seconds)

...