Collect db metrics using postgres_exporter

- Fix checkpoints_timed not found error by using the `--no-collector.stat_bgwriter` flag in the docker compose command
  - This flag can be removed once the issue below is fixed, and a new docker image is released
  - prometheus-community/postgres_exporter#1060
- https://github.com/prometheus-community/postgres_exporter
- https://www.rockdata.net/external/postgres_exporter-metrics/
- https://stackoverflow.com/a/59366582
- https://hub.docker.com/r/prometheuscommunity/postgres-exporter

Author: High5Apps

compose.override.dev.yaml

@@ -9,6 +9,9 @@ services:
     environment:
       SELF_SIGN: true
     restart: "no"
+  metrics-db:
+    environment:
+      DATA_SOURCE_URI: db:5432/organize_development?sslmode=disable
   site:
     build:
       target: dev

compose.yaml

@@ -41,6 +41,17 @@ services:
     restart: always
     volumes:
       - metrics-data:/prometheus:rw
+  metrics-db:
+    command: '--no-collector.stat_bgwriter'
+    depends_on:
+      - db
+    environment:
+      DATA_SOURCE_PASS: ${ORGANIZE_DATABASE_PASSWORD:?err}
+      DATA_SOURCE_URI: db:5432/organize_production?sslmode=disable
+      DATA_SOURCE_USER: postgres
+    image: prometheuscommunity/postgres-exporter:v0.15.0
+    ports:
+      - "127.0.0.1:9187:9187"
   site:
     build:
       context: site

metrics/prometheus.yml

@@ -2,3 +2,6 @@ scrape_configs:
   - job_name: prometheus
     static_configs:
       - targets: ['localhost:9090']
+  - job_name: postgres
+    static_configs:
+      - targets: ['metrics-db:9187']