[📍] Structured output from Telemetry tests for Pinpoint |
|||||
Issue descriptionBackground When a Pinpoint job fails, it posts a comment on the linked bug. Often, users don't click through to the Pinpoint UI to see what happened. Therefore, Pinpoint needs to plumb the error messages through to the user. Currently, Pinpoint parses the logs for Python stack traces. Problem We shouldn't rely on flaky parsing behavior. We also have no robust way to pull out error messages that don't have a stack trace, in cases like a nonexistent benchmark [1] or invalid arguments [2]. Solution We'd like Telemetry to produce an output file containing structured results. Telemetry already produces a GTest-style file listing the passes/failures/skips, but it contains no error details. We can either expand the existing file format or add a new output file type. [1] https://pinpoint-dot-chromeperf.appspot.com/job/14bb6286a40000 [2] https://pinpoint-dot-chromeperf.appspot.com/job/14a85420640000
,
Aug 7
Actually thinking about this more, we probably want to implement this in Telemetry which have all the information about the failure (with https://cs.chromium.org/chromium/src/third_party/catapult/telemetry/telemetry/internal/results/page_test_results.py?rcl=3cb00fbd56816b13ff51170a811a5a2938b32750&l=525) If we implement this in wrapper script, we basically end up doing the log parsing in this layer.
,
Jan 8
,
Jan 8
,
Jan 16
(6 days ago)
,
Jan 16
(6 days ago)
|
|||||
►
Sign in to add a comment |
|||||
Comment 1 by nednguyen@chromium.org
, Aug 6Owner: ----
Status: Available (was: Assigned)