Better error detection/reporting of invalid values in docker-compose?

docker-compose
#1

I recently ran into a problem with running a multi-container app. It was entirely self-inflicted - one of the services was not starting, and ultimately the reason turned out to be a typo in the cap_add section of the service’s definition in docker-compose - I wanted to add a SYS_ADMIN capability, but instead was adding a SYS_ADM one (typed from memory, didn’t check…)
However, the point is that it was quite hard to figure out that this was the reason. The service on the dashboard was just stuck in the Installed state. No logs coming out of the service, clicking start / restart didn’t create any messages either. When checking in the base OS, the container was never running. I eventually realised that the value was wrong when looking at the compose file, but maybe there would be some way of checking for such errors / pointing the user to the reason why the service is not starting? Would be great for those less proficient with Docker!

#3

Hi @maciej.ldc, thank you for the thoughts. Indeed, we need to improve how we show error messages in general.

We have this open issue related to this:

Feel free to subscribe to that issue to get updated as we make progress on this front.

I’ll check with my team if there are other plans about tackling this.