Dashboards are Inconsistent on Docker Compose Restart

Grafana Dashboards
1

Hi, I’m facing a strange issue with dashboard persistence when running Grafana in Docker. My dashboards seem to revert to an older state randomly after I restart my containers.

I will create and save a set of dashboards, and they work perfectly. However, when I stop and restart my stack using docker-compose down and then docker-compose up -d, one of two things happens:

  1. Sometimes, the correct, most recently saved dashboards appear.

  2. Other times, a completely different, older set of dashboards appears, and all my recent work is gone.

This seems to happen unpredictably. If I restart again, the “good” dashboards might reappear. Any advice?

2

Obvious question: so how do you handle persistence? Volume, db, grafana version,… you should provided details.

3

Here is the information about my setup:

  • Grafana Version: I’m using the grafana/grafana:latest Docker image.

  • Persistence Method: I am using the default built-in SQLite database, persisted to a Docker named volume.

  • Deployment: The entire stack is managed with Docker Compose on Windows with WSL2.

docker-compose.yml conf :

services:
grafana:
image: grafana/grafana:latest
container_name: grafana
ports:

  • “3001:3000”
    volumes:
  • grafana-data:/var/lib/grafana
    networks:
  • elastic-stack
    depends_on:
  • prometheus
  • elasticsearch
    restart: unless-stopped

volumes:
grafana-data:

Do you need something else ?

4

That’s only image version. What’s a Grafana version. Don’t assume latest image = latest Grafana version

5

My Grafana version is v12.0.2.

6

That SQLite is pure fs approach, so play it with that. Check some fsync options on your os and used docker storage driver, maybe sqllite level. I would try proper DB (postgres) they works better with with fs usually.

1 Like
7

thank you for the advice