Closed
Bug 1461482
Opened 6 years ago
Closed 6 years ago
Verify VCS Perfherder data for Windows tasks is being ingested
Categories
(Testing Graveyard :: ActiveData, defect, P1)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: gps, Assigned: ekyle)
References
Details
I'm able to run ActiveData queries against the VCS Perfherder metrics being recorded by `hg robustcheckout`. However, I'm selecting the raw data and am not seeing any data from Windows workers. There's a chance I'm doing it wrong. But some of the aggregate data seems to reflect no data for Windows. https://taskcluster-artifacts.net/Yl5tnpaBQTGjduaj76r2PQ/0/public/logs/live_backing.log is an example of a Windows task whose data is not ingested. Search for "PERFHERDER_DATA". The fact it is at the beginning of a line may confuse the log parser?
Reporter | ||
Comment 1•6 years ago
|
||
FWIW, I have other patches in flight that attempt to make Firefox CI tasks use run-task. At that point, logs will be prefixed with the typical time prefix and this may cause ActiveData to start recognizing the PERFHERDER_DATA lines. So if this is a non-trivial amount of work, ping me before you invest much time and I may be able to save you the work!
Assignee | ||
Updated•6 years ago
|
Flags: needinfo?(klahnakoski)
Assignee | ||
Comment 2•6 years ago
|
||
Sorry, I never even saw this bug. There is a dup somewhere, which I also did nothing on. needinfo myself so I remember to act on this.
Flags: needinfo?(klahnakoski)
Assignee | ||
Updated•6 years ago
|
Flags: needinfo?(klahnakoski)
Assignee | ||
Comment 4•6 years ago
|
||
Flags: needinfo?(klahnakoski)
Assignee | ||
Comment 5•6 years ago
|
||
Turns out it was two problems. The multiline json (with taskcluster timestamps)
[taskcluster 2019-03-28T17:35:29.768Z]
and the PERFHERDER record has no timestamp
Here is the file
https://taskcluster-artifacts.net/Qu-_x508RWys08fcgMtVgQ/0/public/logs/live_backing.log
Here is he task record
{
"from":"task",
"select":"_id",
"where":{"eq":{"task.id":"Qu-_x508RWys08fcgMtVgQ"}}
}
Here are the (new) perf records
{
"from":"perf",
"select":["_id","run.key","result.framework.name"],
"where":{"and":[{"eq":{"task.id":"Qu-_x508RWys08fcgMtVgQ"}}]},
"limit":1000
}
Processing is proceeding atarting a few hours back.
Assignee | ||
Updated•6 years ago
|
Flags: needinfo?(klahnakoski)
Assignee | ||
Updated•6 years ago
|
Status: NEW → RESOLVED
Closed: 6 years ago
Flags: needinfo?(klahnakoski)
Resolution: --- → FIXED
Updated•3 years ago
|
Product: Testing → Testing Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•