Update Jaeger CRD, add note

Update to the latest for 1.24.0: https://github.com/jaegertracing/jaeger-operator/blob/v1.24.0/deploy/crds/jaegertracing.io_jaegers_crd.yaml

Add a note in the readme about why we maintain the CRD different from the upstream chart.

changed iteration to Big Bang Iterations Jul 13, 2021 - Jul 26, 2021

@runyontr the crd for jaeger is quite different from upstream. It looked like this was to provide validation that we are passing a correctly formatted spec but I wasn't sure.

added statusdoing label

set weight to 1

After a quick chat with Tom, the helm chart does not properly maintain the CRD to match the upstream one. Will be updating the CRD and adding a note about it to our README.

changed title from ask Tom why the CRD is the way it is to Update Jaeger CRD, add note

changed the description

created merge request !12 (merged) to address this issue

mentioned in merge request !12 (merged)

set weight to 2

Investigating the pipeline failure...also need to add a job/subchart to handle the CRD updates.

changed milestone to %1.13.0

added jaeger + 1 deleted label

Got CRD updates handled by an upgrade job today. Last step will be debugging the pipeline/fresh installs.

It seems like the CRD changed and does not support IPS at all the same spots we had it before. This may take some investigation since it seems to intermittently work...

Hacked up the CRD today and added the missing fields: https://repo1.dso.mil/platform-one/big-bang/apps/core/jaeger/-/merge_requests/12/diffs?commit_id=89ba422d7c2888478e5fcc340218c9b8dcc4beba

This seems to solve the CI and clean install issue but I'm not sure why and don't like it. I'm not sure if there's more validation going on/required for this new CRD or something and the fields we set before were never supposed to work? Investigating further tomorrow...

We were apparently setting values that didn't match the CRD in the past. I've removed those and things seem happy. Needs some more investigation to make sure all images pull properly still in all scenarios. Upstream issues/PRs seem to indicate the use of a serviceaccount is required for IPS on all components except the agent sidecar, which requires setting them in the spec directly. In practice everything pulls fine on an all-in-one install, but also need to test the prod config with separate pods for each.

added statusreview label and removed statusdoing label

added statusdoing label and removed statusreview label

closed with merge request big-bang/bigbang!689 (merged)

removed statusdoing label