1503380 - When running raptor gecko profiling jobs, raptor should still save results to a raptor.json artifact

Status: RESOLVED FIXED

(Testing :: Raptor, enhancement)

Description

[Robert Wood [:rwood]]

Currently when running Raptor in production with gecko profiling (--gecko-profile) turned on, we don't write the results to a .json artifact (since we don't dump out PERFHERDER_DATA formatted results when running with profiling as we don't want perfherder to ingest the measurements).

We should still write the results for the 2 profiling pagecycles to a raptor.json anyway and have it saved as a taskcluster artifact, so that when developers are looking at the profiles they can also see the actual raw results.

When running locally with --gecko-profile the results are available in /testing/mozharness/build/raptor.json - but such a .json is not available on production profiling runs.

Comment 1

[Robert Wood [:rwood]]

Attachment: Bug 1503380 - When running raptor gecko profiling jobs, raptor should still save results to a raptor.json artifact; r?jmaher

Comment 2

[Robert Wood [:rwood]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=bc589e588ab9e00d6a63fae3f9c527356472f49a

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/c06ea2a391cb
When running raptor gecko profiling jobs, raptor should still save results to a raptor.json artifact; r=jmaher

Comment 4

[Andrei Ciure[:aciure]]

https://hg.mozilla.org/mozilla-central/rev/c06ea2a391cb