We have been asked by the host to reduce the load on the CI.
So I have already proposed to skip pipeline for draft merge request so use the draft mark when possible.
It would also help to have more runners for our own usage.
Another lead would be to reduce the matrix for python versions and backends. For example we actually just need to run code for each python version once and for each backend once.
Does anyone have an idea how to improve this?
I can provide one more, send me the details needed to configure one.
Do we know how many runners we need to add?
You just have to follow Run GitLab Runner in a container | GitLab but use the heptapod runner.
I will send you the register token.
I would like to help.
Do you need any special description for the runner?, maybe an email adress or heptapod username.
Not really, I just sent the token to register the runner.
Indeed we have this kind of convention to name the runner: <organisation>-<name or id>-<executor (docker)>-<arch>.
Ex: tryton-tycho-docker-amd64
I created Test last Python version with PostgreSQL (!1995) · Merge requests · Tryton / Tryton · GitLab
and the typical ressources needed for a runner ?
Not sure but probably at least one CPU and 500Mb
Your runner is down since 2 weeks and never run any jobs.
So I guess you are no more proposing it. I will remove it in few days.
I will check tonight, I was not conscious of the down. I’m still open to contribute the runner.