Health-check another container

[opal stump]

Hi friends. I'm new to Dagger, be gentle. I've got a dev container (TiDB) which is spontaneously hanging--or so I suspect. I routinely notice that I can no longer connect to it or run queries, and by bringing it down and up again everything is fine. It's killing my vibe a little 🙂

Now, on the one hand, I'd love to learn how to get richer diagnostics and see if I'm taking crazy pills and/or what's causing this failure mode.

On the other hand, I'm exploring a passive health-check, which can be done with a shell script that runs SELECT 1 every few seconds, but wondering if it makes sense to stand up another dagger container to monitor the misbehaving one. RFC: Would that be a weird use-case, an antipattern, or somewhat reasonable?

[stark shale]

We're always gentle 🙂

Quick clarifying question, is the hanging container run by Dagger? Or external.

Either way it seems like a reasonable use case

[stark shale]

if it's external: dagger core container from --adress=mysql with-service-binding --endpoint=tcp://localhost:4000 --alias=tidb with-exec --args=sh,-c,'while true; do mysql -h tidb:4000 -e "select 1"; done'

[opal stump]

Run by dagger yes, not external. Thanks Solomon!

[stark shale]

in that case just pass myTiDBContainer.asService() as argument to the service binding and you should be good

[barren folio]

👋 welcome! ask as many questions as you need, we'll try our best to help out.

First and foremost, could you describe what setup (os version, dagger version and container runtime) you're currently using?

How are you currently connecting to that TiDB container? Is the TiDB spawned a dagerr call tidb up? Or you're using TiDB as a dependency of some other app?