Unable to see Event in about:telemetry in artifact build
Categories
(Toolkit :: Telemetry, defect, P1)
Tracking
()
People
(Reporter: ewright, Assigned: janerik)
References
Details
Attachments
(1 file)
I was recently adding some event telemetry using Events.yaml. The event was in the content process.
- I ran the build in artifact mode
- I visited about:telemetry and could see the event show up in "keyed scalars > dynamic" but I saw no events listed under "Events".
- I ran the build in regular mode
- I visited about:telemetry again and saw the event show under "events > content"
I was under the impression that an artifact build could be used for this.
Comment 1•4 years ago
|
||
Definitely the artifact build should work exactly the same as a full build in this respect. Thank you for filing this bug.
Seems like there's a hole in our Build Faster support for Events, maybe related to bug 1563239
Updated•4 years ago
|
Comment 2•4 years ago
|
||
I'm having difficulty reproducing this with latest code. I tried
- Changing
.mozconfig
to build artifact - Adding a new category and event definition to
Events.yaml
./mach build faster && ./mach run
- Loading
about:telemetry
and opening the dev console toTelemetry.setEventRecordingEnabled
andTelemetry.recordEvent
- Refreshing
about:telemetry
and checking the Events section
The event shows up properly, and is even summarized properly to Keyed Scalars > Parent. I must be missing something.
:ewright Can you help me find out what I'm missing?
Reporter | ||
Comment 3•4 years ago
|
||
(In reply to Chris H-C :chutten from comment #2)
I'm having difficulty reproducing this with latest code. I tried
- Changing
.mozconfig
to build artifact- Adding a new category and event definition to
Events.yaml
./mach build faster && ./mach run
- Loading
about:telemetry
and opening the dev console toTelemetry.setEventRecordingEnabled
andTelemetry.recordEvent
- Refreshing
about:telemetry
and checking the Events sectionThe event shows up properly, and is even summarized properly to Keyed Scalars > Parent. I must be missing something.
:ewright Can you help me find out what I'm missing?
sorry for taking a while, I was on PTO. I'll see if I can duplicate it again.
Comment 4•3 years ago
|
||
Did it happen to happen again? Or shall I close this down?
Comment 5•3 years ago
|
||
Closing for now.
Reporter | ||
Comment 6•3 years ago
|
||
:chutten, thank you, sorry I haven't had time try to reproduce this yet. I'll re-open if I can reproduce again.
Updated•3 years ago
|
Comment 7•3 years ago
|
||
I can reproduce this. In both artifact and non-artifact builds, EventArtifactDefinitions.json next to my locally built binary is not json - it's a c++ header file - ie, it starts with:
/* This file is auto-generated, only for internal use in TelemetryEvent.h,
see gen_event_data.py. */
#ifndef mozilla_TelemetryEventData_h
#define mozilla_TelemetryEventData_h
#include "core/EventInfo.h"
namespace {
#if defined(_MSC_VER) && !defined(__clang__)
const uint32_t gExtraKeysTable[] = {
#else
constexpr uint32_t gExtraKeysTable[] = {
#endif
The key different between artifact and non-artifact builds is that presumably the latter always has up-to-date events builtin so doesn't care about loading dynamic ones at runtime.
The console always reports:
1574227665947 Toolkit.Telemetry ERROR TelemetryController::registerEventProbes - there was an error loading EventArtifactDefinitions.json: SyntaxError: JSON.parse: unexpected character at line 1 column 1 of the JSON data(resource://gre/modules/TelemetryController.jsm:1355:28) JS Stack trace: registerEventProbes@TelemetryController.jsm:1355:28
async*registerJsProbes@TelemetryController.jsm:1265:16
async*setupTelemetry@TelemetryController.jsm:740:43
observe@TelemetryController.jsm:1001:21
observe@TelemetryController.jsm:187:17
TelemetryStartup.prototype.observe@TelemetryStartup.jsm:31:25
which is a manifestation of the above. Note that after a 'build faster', this file does have the new probes - ie, that file is successfully getting rebuild when necessary - it's just the wrong file!
I'm on Windows which tends to be rare around here, so I wonder if that has anything to with the difficulty in others reproducing?
Assignee | ||
Updated•3 years ago
|
Assignee | ||
Updated•3 years ago
|
Assignee | ||
Comment 8•3 years ago
|
||
Looks like this was mistakenly introduced here:
https://phabricator.services.mozilla.com/D48409#change-MERCC5y13iOg
Pushed by jrediger@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/77735f86efb6 Fix generation of event artifact definitions caused by using the wrong entry point r=chutten
Comment 10•3 years ago
|
||
bugherder |
Comment 11•3 years ago
|
||
Since the status are different for nightly and release, what's the status for beta?
For more information, please visit auto_nag documentation.
Updated•3 years ago
|
Description
•