Follow

I have a VM which runs a bunch of docker containers. Now and then it becomes unresponsive with CPU and disk usage hitting the peak. The only way to interact after that is to restart the VM. What can I use to identify the offending program?

@tecoholic

Just docker or kubernetes ? If it is kubernetes there are plenty of debugging options `kubectl top pods`; `prometheus` etc.

One simple thing that would help is to pass cpu and memory limits which would restrict and not make the machine shutdown.

@psankar No Kubernetes. Just a plain VM with docker containers running in them

@psankar Turns out it can be used only for Docker Swarm Mode. Have setup a CRON job with iotop to log every minute.

@psankar For direct "docker run"s that's okay. But for "docker-compose up" runs the parameters are not considered unless it's on swarm mode.

@tecoholic Are you sure ? docs.docker.com/compose/compos does not mention anything like that. Search for "Resources" in that page. If so, this must be documented.

@tecoholic Have you considered a monitoring solution like ? You can use it for free

@koalalorenzo Haven't known about it. I will look into it. Thank you for the tip

Sign in to participate in the conversation
Mastodon

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!