Monitoring Docker Swarm task rejection

20 views
Skip to first unread message

Chan Lee

unread,
Oct 23, 2020, 5:00:32 PM10/23/20
to Prometheus Users
Hi all,

When monitoring a Docker Swarm, is there a way to track task rejection?

We're currently tracking container health by looking at the change in start time but this doesn't cover the case where a task is rejected (no start time is recorded as it never starts). In our case, a task was continually being rejected because the service was trying mount a directory that didn't exist.

I've done some searching around but I can't seem to find either a query or an exporter that would fill this need. Are there any solutions that you all know about?

We are currently using the following exporters:

Julien Pivotto

unread,
Oct 23, 2020, 5:38:50 PM10/23/20
to Chan Lee, Prometheus Users
On 23 Oct 14:00, Chan Lee wrote:
> Hi all,
>
> When monitoring a Docker Swarm, is there a way to track task rejection?
>
> We're currently tracking container health by looking at the change in start
> time but this doesn't cover the case where a task is rejected (no start
> time is recorded as it never starts). In our case, a task was continually
> being rejected because the service was trying mount a directory that didn't
> exist.
>
> I've done some searching around but I can't seem to find either a query or
> an exporter that would fill this need. Are there any solutions that you all
> know about?
>
> We are currently using the following exporters:
>
> - cAdvisor
> - Node Exporter
> - https://github.com/prometheus-net/docker_exporter

Are you using the docker swarm discovery ? I would expect that the task
should be scraped anyway ( and fail )


>
>
> --
> You received this message because you are subscribed to the Google Groups "Prometheus Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-use...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/773e125e-d7f4-4bb6-b117-2ec217a7ce3an%40googlegroups.com.


--
Julien Pivotto
@roidelapluie

Chan Lee

unread,
Oct 26, 2020, 3:24:57 PM10/26/20
to Prometheus Users
Yes, we're using it.

That's the strange thing, I can see it via CLI and through Portainer, but it doesn't register as a new container in the exported metrics.

Reply all
Reply to author
Forward
0 new messages