Closed Bug 874175 Opened 13 years ago Closed 12 years ago

Add Telemetry probe for ICE completion success rate

Categories

(Core :: WebRTC, defect)

defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla26

People

(Reporter: derf, Assigned: jib)

References

Details

(Whiteboard: [WebRTC][blocking-webrtc-])

Attachments

(1 file, 3 obsolete files)

As a first stab at adding telemtry, we should add a probe to count the number of successfully established ICE connections vs. the number attempted.
Assignee: nobody → ben
Blocks: 868626
Whiteboard: [WebRTC],[blocking-webrtc-]
Assignee: ben → bbrittain
No longer blocks: 868626
Whiteboard: [WebRTC],[blocking-webrtc-]
Blocks: 868626
Whiteboard: [WebRTC][blocking-webrtc-]
Assignee: bbrittain → ben
Attachment #752398 - Flags: review?(tterribe)
Attachment #752398 - Flags: review?(nfroyd)
(In reply to Ben Brittain (:bbrittain) from comment #1) > Created attachment 752398 [details] [diff] [review] Pushed to try: https://tbpl.mozilla.org/?tree=Try&rev=e8f8db977c16
Depends on: 874670
No longer depends on: 874670
Comment on attachment 752398 [details] [diff] [review] Adds telemetry for ice success rates Review of attachment 752398 [details] [diff] [review]: ----------------------------------------------------------------- r-, but it should be easy to fix this up. ::: dom/media/PeerConnection.js @@ +934,5 @@ > // we go ahead and trigger a call of onicecandidate here. > // This is to provide some level of compatibility with > // scripts that expect this behavior (which is how Chrome > // signals that no further trickle candidates will be sent). > + // TODO: This needs to be removed when Bug 842459 lands. Bug 842549. @@ +935,5 @@ > // This is to provide some level of compatibility with > // scripts that expect this behavior (which is how Chrome > // signals that no further trickle candidates will be sent). > + // TODO: This needs to be removed when Bug 842459 lands. > + // TODO: Fix that bug number! That probably shouldn't be in this patch, though, as it's unrelated to this change. Feel free to upload a separate patch to bug 856425, which added this comment, to fix this. @@ +998,5 @@ > this.callCB(this._ongatheringchange, "gathering"); > break; > case Ci.IPeerConnection.kIceConnected: > // ICE gathering complete. > + histogram.add(0); IF WEBRTC_ICE_SUCCESS_RATE has kind "boolean", I think the value you add should be true/false. In which case completion should probably be true, and failure false. ::: toolkit/components/telemetry/Histograms.json @@ +3313,5 @@ > "description": "The time (in milliseconds) that it took a 'detach' request to go round trip." > + }, > + "WEBRTC_ICE_SUCCESS_RATE":{ > + "kind": "boolean", > + "description": "Number of sent ICE connections vs sucessful ICE connections" What you're actually measuring is successful ICE connections vs. failed ICE connections. Also, fix the spelling of "successful" and punctuate properly (periods after "vs" and at the end of the phrase).
Attachment #752398 - Flags: review?(tterribe) → review-
Attachment #752818 - Flags: review?(tterribe)
contrary to the name of the boolean telemetry type, it takes 0/1 not false/true
Comment on attachment 752398 [details] [diff] [review] Adds telemetry for ice success rates Review of attachment 752398 [details] [diff] [review]: ----------------------------------------------------------------- I have nothing to add to Timothy's review, other than to note that histogram.add() for a boolean histogram will work just fine with a numeric or boolean arg. A boolean arg is generally clearer and IMHO better in this case.
Attachment #752398 - Flags: review?(nfroyd)
Attachment #752818 - Attachment is patch: true
Attachment #752398 - Attachment is obsolete: true
Comment on attachment 752818 [details] [diff] [review] Adds telemetry for ice success rates; patch2 Review of attachment 752818 [details] [diff] [review]: ----------------------------------------------------------------- r=me with the one extraneous change removed. I'll leave it up to you whether you want to use 0/1 or true/false, since froydnj says they both work. Don't forget to add a "r=derf" to the end of the checkin message before asking for checkin. ::: dom/media/PeerConnection.js @@ +934,5 @@ > // we go ahead and trigger a call of onicecandidate here. > // This is to provide some level of compatibility with > // scripts that expect this behavior (which is how Chrome > // signals that no further trickle candidates will be sent). > + // TODO: This needs to be removed when Bug 842459 lands. Please don't add trailing whitespace here.
Attachment #752818 - Flags: review?(tterribe) → review+
(In reply to Ben Brittain (:bbrittain) from comment #5) > contrary to the name of the boolean telemetry type, it takes 0/1 not > false/true Where do you see this? I just tested in-browser to confirm and the .add method does accept true/false in addition to 0/1.
I had originally tried true/false and not had it work. It must have been an unrelated problem. sorry about the confusion.
Attachment #752818 - Attachment is obsolete: true
Attachment #752854 - Flags: review?(tterribe)
Attachment #752854 - Attachment is patch: true
Comment on attachment 752854 [details] [diff] [review] Adds telemetry for ice success rates; r=derf Review of attachment 752854 [details] [diff] [review]: ----------------------------------------------------------------- r=me again, but don't forget to add a commit message with the bug number and reviewer. See <https://developer.mozilla.org/en-US/docs/Mercurial_FAQ#How_can_I_generate_a_patch_for_somebody_else_to_check-in_for_me.3F>. ::: toolkit/components/telemetry/Histograms.json @@ +3310,5 @@ > "kind": "exponential", > "high": "10000", > "n_buckets": "1000", > "description": "The time (in milliseconds) that it took a 'detach' request to go round trip." > + }, Note that this will require rebasing.
Attachment #752854 - Flags: review?(tterribe) → review+
Reassigning to Jan-Ivar since Ben has gone back to school.
Assignee: ben → jib
Target Milestone: --- → mozilla26
Rebased and verified that new stat appears in about:telemetry. Carrying forward r+ from derf.
Attachment #795561 - Flags: review+
Attachment #752854 - Attachment is obsolete: true
Blocks: 874670
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: