The https://wiki.jenkins-ci.org/display/JENKINS/Support+Core+Plugin automatically tracks slow requests (eg if a request takes > 10s, it will take threadumps and do other diagnostic things). The benefit is that when someone is having a bad time, the evidence is already recorded to point to the culprit.
In a SPA like blue ocean, this works nicely for the REST endpoints (which are a source of problems, but not alone). What we lose is the ability to map slowness in the support bundle to what the user is seeing.
- Investigate if there are ways to track time taken by react components (render time?)
- Investigate if there are ways to track time used in routes (render time?)
- Collect profiling data, and publish to (new) endpoint