siren-tests: "sirenada-stg-t3-micro" job appears in Argus as 'Not Planned' although it runs on a daily basis.

[rayakurl]

https://jenkins.scylladb.com/job/siren-tests/job/Weekly_Runs/job/sirenada-stg-t3-micro/ appears in Argus as 'Not Planned' although it runs on a daily basis.

[k0machi]

This is due to the fact that it exits too early, so there are no results, we should make it so that the upload script still submits an empty test to indicate that it happened.

[fruch]

@k0machi

can you explain that is the need changed needed on siren-test pipelines ? (and reference the change @temichus did in SCT)

[k0machi]

@k0machi

can you explain that is the need changed needed on siren-test pipelines ? (and reference the change @temichus did in SCT)

Yes, same deal as SCT we need to make similar plugin API that we have there where a test is first created and then populated with data, so that early failures are not hidden from us.

[fruch]

@k0machi

can you explain that is the need changed needed on siren-test pipelines ? (and reference the change @temichus did in SCT)

Yes, same deal as SCT we need to make similar plugin API that we have there where a test is first created and then populated with data, so that early failures are not hidden from us.

And again can you estimate how much work it is ?

[k0machi]

And again can you estimate how much work it is ?

I'd say 3 days or so, need to both update argus, its client and pipeline files.

[roydahan]

The time estimation field is 8 hours, but you say 3 days.
Any workaround we can do here?

Can I see a screenshot of the issue / link to such job?

[k0machi]

The issue here is we don't have a stage in those pipelines like we do in SCT that pre-creates argus run when pipeline starts, so currently those pipelines will show nothing in case of early failure, e.g. spot termination (as siren tests submit everything at the end of the run)

[fruch]

The issue here is we don't have a stage in those pipelines like we do in SCT that pre-creates argus run when pipeline starts, so currently those pipelines will show nothing in case of early failure, e.g. spot termination (as siren tests submit everything at the end of the run)

Can you link and point to how such a stage can be added, so we can move this into siren-tests

[k0machi]

The issue here is we don't have a stage in those pipelines like we do in SCT that pre-creates argus run when pipeline starts, so currently those pipelines will show nothing in case of early failure, e.g. spot termination (as siren tests submit everything at the end of the run)

Can you link and point to how such a stage can be added, so we can move this into siren-tests

I'll do a write up once #289 is complete with similar steps that need to be done there.