zerotier-app starts before zerotier-db is ready, causing database connection errors

[DeepseaGoldenFlow]

Hi,

I'm running the tiredofit/zerotier Docker image with Docker Compose on Windows 10 using Docker Desktop with WSL2 backend.
I noticed that when I do docker-compose up -d, the zerotier-app container starts before the zerotier-db (PostgreSQL) service is fully ready, causing repeated connection errors like:

"[WARN] Postgres Host 'zerotier-db' is not accessible, retrying..."

This happens because Docker Compose's depends_on only waits for container start, not for service readiness.

To fix this, I suggest adding a healthcheck to the zerotier-db service (using pg_isready) and configuring depends_on.condition in the compose file, so that zerotier-app waits for the database to be healthy before starting.

This would improve startup reliability and reduce confusing log entries.

Thanks for your great work on this project!

Best regards

[tiredofit]

That's actually my db_ready function that occurs within the tiredofit/alpine base image. Which is what the tiredofit/nginx and then this tiredofit/zerotier image is based on. It's a function that handles some basic connectivity "is it up" before I perform "is it writeable" checks. It works outside of the compose ecosystem for other orchestrators, and as well predates some of the healthcheck stuff that came with later compose specifications. Some of these images have history of almost 9 years! That being said - I have a new set of base images that I've been slowly working on over the course of a couple years that solve some of those visual annoyances.
Glad the image is working out for you :)