If you think a bug might affect users in the 57 release, please set the correct tracking and status flags for Release Management.

Intermittent histogram has expected content - [888,157,754] deepEqual [0,0,0,0,0,0] - JS frame :: chrome://mochitests/content/browser/browser/components/search/test/browser_aboutSearchReset.js :: checkTelemetryRecords :: line 26

RESOLVED FIXED in Firefox 53

Status

()

Toolkit
Telemetry
--
blocker
RESOLVED FIXED
10 months ago
2 months ago

People

(Reporter: Treeherder Bug Filer, Assigned: florian)

Tracking

({intermittent-failure})

unspecified
mozilla53
intermittent-failure
Points:
---

Firefox Tracking Flags

(firefox50 unaffected, firefox51 unaffected, firefox52 unaffected, firefox53 fixed)

Details

Attachments

(1 attachment)

(Reporter)

Description

10 months ago
treeherder
Filed by: tomcat [at] mozilla.com

https://treeherder.mozilla.org/logviewer.html#?job_id=5591259&repo=mozilla-central

https://queue.taskcluster.net/v1/task/RY5t1sXHSruQAdq6xlbtRw/runs/0/artifacts/public%2Flogs%2Flive_backing.log

Comment 1

10 months ago
i guess this something for you guys
Flags: needinfo?(gfritzsche)
Flags: needinfo?(alessio.placitelli)

Comment 2

10 months ago
actually now that we merge day this expired      "expires_in_version": "53",

florian can you fix this asap since this is hitting m-c
Flags: needinfo?(gfritzsche)
Flags: needinfo?(florian)
Flags: needinfo?(alessio.placitelli)

Comment 3

10 months ago
(In reply to Carsten Book [:Tomcat] from comment #2)
> actually now that we merge day this expired      "expires_in_version": "53",
> 
> florian can you fix this asap since this is hitting m-c

btw let me when you have a patch etc and i can direct land this on m-c

Comment 4

10 months ago
and btw its closing m-c tree
Severity: normal → blocker
(Assignee)

Comment 5

10 months ago
Created attachment 8810473 [details] [diff] [review]
Keep the SEARCH_RESET_RESULT Telemetry probe until Firefox 57, r=

Shipping the search reset feature has been delayed as we want to test it with a Shield experiment before deploying to all the release population, so we very much need this probe for another couple months. Bumping to 57 to add 4 releases / about 6 months.

I find it strange that the test fails with the histogram containing random values, rather than an exception saying the histogram doesn't exist, but I verified locally that bumping the version limit on the probe is all it takes to fix this.
Attachment #8810473 - Flags: review?(benjamin)
(Assignee)

Updated

10 months ago
Assignee: nobody → florian
Status: NEW → ASSIGNED

Updated

10 months ago
Attachment #8810473 - Flags: review?(benjamin) → review+

Comment 6

10 months ago
Pushed by ryanvm@gmail.com:
https://hg.mozilla.org/mozilla-central/rev/71fd23fa0803
Keep the SEARCH_RESET_RESULT Telemetry probe until Firefox 57. r=bsmedberg
Status: ASSIGNED → RESOLVED
Last Resolved: 10 months ago
status-firefox53: --- → fixed
Flags: needinfo?(florian)
Resolution: --- → FIXED
Target Milestone: --- → mozilla53
> I find it strange that the test fails with the histogram containing random
> values, rather than an exception saying the histogram doesn't exist, but I

Should a followup bug be filed on that issue?
(Assignee)

Comment 8

10 months ago
(In reply to Bill Gianopoulos [:WG9s] from comment #7)
> > I find it strange that the test fails with the histogram containing random
> > values, rather than an exception saying the histogram doesn't exist, but I
> 
> Should a followup bug be filed on that issue?

Is it the expected behavior that reading data from an expired histogram using Services.telemetry.getHistogramById(name).snapshot() returns odd values? I would have expected that to throw an exception.

Hopefully Georg knows.
Flags: needinfo?(gfritzsche)

Comment 9

10 months ago
There is a bug tree at bug 1311100 for addressing semantic weirdness of the Telemetry APIs in exceptional cases. This one in particular isn't called out, so I'll wait for :gfritzsche's ni? then file a bug :)
> Is it the expected behavior that reading data from an expired histogram
> using Services.telemetry.getHistogramById(name).snapshot() returns odd
> values? I would have expected that to throw an exception.

As opposed to even values?  (Sorry I could not resist) ;-)
(In reply to Chris H-C :chutten from comment #9)
> There is a bug tree at bug 1311100 for addressing semantic weirdness of the
> Telemetry APIs in exceptional cases. This one in particular isn't called
> out, so I'll wait for :gfritzsche's ni? then file a bug :)

This is a quirk of the histogram expiry implementation. (all expired histograms are treated as the same one histogram).
I filed bug 1317702 about this.
Flags: needinfo?(gfritzsche)
status-firefox50: --- → unaffected
status-firefox51: --- → unaffected
status-firefox52: --- → unaffected

Comment 12

10 months ago
18 failures in 715 pushes (0.025 failures/push) were associated with this bug in the last 7 days.  

Repository breakdown:
* try: 13
* mozilla-central: 5

Platform breakdown:
* linux64: 14
* windows8-64: 2
* linux32: 2

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1317296&startday=2016-11-14&endday=2016-11-20&tree=all
Depends on: 1383770
You need to log in before you can comment on or make changes to this bug.