This is probably beyond me as a Python dabbler, but I think it may be nice to make the check more flexible in how --check-all behaves.
Currently if a statically defined check does not have a matching GitLab health API subject, that check is flagged w/ UNKNOWN status as a monitoring plugin should do when it cannot parse output.
My proposal would be to either by default, or as an option setting (eg --ignore-missing), allow --check-all to skip any checks that don't map to a health API subject.
Perhaps as part of the 'OK' response, it could also return something like 'API doesn't include checks foo, and bar'. And to help flag for future health API changes, it could also return something like 'API includes unmonitored subject foo' when it discovers a health subject that we don't already check for.
This is probably beyond me as a Python dabbler, but I think it may be nice to make the check more flexible in how
--check-allbehaves.Currently if a statically defined check does not have a matching GitLab health API subject, that check is flagged w/ UNKNOWN status as a monitoring plugin should do when it cannot parse output.
My proposal would be to either by default, or as an option setting (eg
--ignore-missing), allow--check-allto skip any checks that don't map to a health API subject.Perhaps as part of the 'OK' response, it could also return something like 'API doesn't include checks foo, and bar'. And to help flag for future health API changes, it could also return something like 'API includes unmonitored subject foo' when it discovers a health subject that we don't already check for.