Concept

Some critical information about pigsty

Pigsty is logically composed of two parts: monitoring system and provisioning solution.

The monitoring system is responsible for monitoring the PostgreSQL database cluster, and the provisioning solution is responsible for creating the PostgreSQL database cluster. Before understanding Pigsty’s monitoring system and provisioning solution, it is helpful to read naming principles and overall architecture to get an intuitive picture of the overall design.

Pigsty’s monitoring system and provisioning solution can be used independently, and users can use Pigsty monitoring system to monitor existing PostgreSQL clusters and instances without using Pigsty provisioning solution, check Monitor-only deployment.

Monitoring System

You can’t manage what you don’t measure.

Pigsty delivers the best open source PostgreSQL monitoring system.

Pigsty’s monitoring system is physically divided into two parts.

server: deployed on meta-node, including services such as Prometheus, a timing database, Grafana, a monitoring dashboard, Altermanager, and Consul, a service discovery.
Client: deployed on [Database Node](arch/#Database Node), including NodeExporter, PgExporter, Haproxy. passively accept Prometheus pull, on.

The core concepts of the Pigsty monitoring system are as follows.

observability
monitoring hierarchy
Identity management
monitoring metrics
[monitoring panels](. /monitor/links/)
alerting rules

Provisioning Solution

It is better to teach a man how to fish rather than give him a fish

Provisioning Solution (Provisioning Solution) refers to a system that delivers database services and monitoring systems to users. The Provisioning Solution is not a database, but a database factory**, where the user submits a configuration to the provisioning system, and the provisioning system creates the required database cluster in the environment according to the user’s required specifications, similar to creating the various resources required by the system by submitting a YAML file to Kubernetes.

Pigsty’s provisioning solution is divided into two parts for deployment.

Infrastructure (Infra) : Deployed on meta nodes to monitor key services such as infrastructure, DNS, NTP, DCS, local sources, etc.
[Database Cluster](arch/# Database Cluster) (PgSQL) : Deployed on database nodes to provide database services to the outside world as a cluster.

Pigsty’s provisioning scheme is deployed to two types of objects.

[meta node](arch/# meta node) (Meta): deployment infrastructure to perform control logic, each Pigsty deployment requires at least one meta node, which can be reused as a normal node.
[Database Node](arch/#Database Node) (Node): used to deploy database clusters/instances, Pigsty uses exclusive deployment with one-to-one correspondence between nodes and database instances.

The relevant concepts of Pigsty provisioning solution are as follows.

Concept

Monitoring System

Provisioning Solution

Naming Rules

Architecture

Monitoring

Provisioning