#Custom Gitlab runners
1 messages · Page 1 of 1 (latest)
By “rolling your own”, do you mean…
-
that you are using self-hosted gitlab runners but would prefer to have runners hosted by gitlab.com?
-
or that you are using self-hosted runners with a forked configuration, and would prefer to run self-hosted with an unmodified official configuration?
Or to put it differently; what would the ideal solution look like for you?
rolling our own means that we need to setup our own team dedicated runners which adds maintenance burden. We have onprem Gitlab and ideal would be that we could use the shared runners of enterprise (managed by other team) but we are not allowed to have dind setup there.
We looked into running the Dagger Engine in Kubernetes but the feature seemed to be quite experimental and poorly documented still.
If Dagger provided a kubernetes configuration to augment your shared gitlab runners with dagger engine pods, would that be useful?. The benefit over a completely separate dagger engine deployment, is that you wouldn’t have to worry about scaling and load balancing in two places: each gitlab runner would get its own dagger engine, but without dind
cc @soft wagon @robust smelt @full cedar
Maybe we could use dagger engine in a rootless mode?
@full cedar since it seems there are several possible options, maybe a good next step could be to list them in the issue, and start comparing pros and cons? https://github.com/dagger/dagger/issues/4515