search

opensafely-core / sysadmin

Various scripts and tools for administering OpenSAFELY organisation and infrastructure
0 stars 0 forks source link

Measure the number of output checking requests #148

Closed lucyb closed 8 months ago

lucyb commented 8 months ago

An output checking request produces a GitHub issue. Can we track these in Grafana as a way of measuring how busy the output checking service is?

lucyb commented 8 months ago

I've started to measure the number of issues in the main output checking repo here https://dashboards.opensafely.org/d/adb76f4d-0bcc-4d0d-adb3-e2fcbfcc6993/product-dashboard-spike?orgId=1

It's currently limited, as we've no way of telling which project or user they belong to, or even whether they're internal or external projects.

Additionally, LSHTM and Bristol do their own output checking and that's not measured here. I don't think that's too much of a problem, since our main aim is to understand how busy our own output checking team is.

We could potentially get more information by instrumenting the use of osrelease. That would give us more information about who is doing output checking and how often files are being released, which could give us an indication of how busy the output checking service is. It might also be able to tell us which workspace (although probably not the project).

lucyb commented 8 months ago

After speaking with Louis, I've been able to create an output checking board that shows the numbers of requests per week for BI, Bristol and LSHTM. The BI ones are broken down by internal/external requests. I've also shown a breakdown of the number of requests received by output checkers.