Bug 1837230 Comment 4 Edit History

Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.

Nope. Another dead end. Only 10 clients report having any queue overflow. Might be because an inability to deliver the "events" ping might influence the ability to deliver the "metrics" ping as well.

To sum:
* We have a long list of things it isn't: 
    * 
    * 
    * 
    *
Nope. Another dead end. Only 10 clients report having any queue overflow. Might be because an inability to deliver the "events" ping might influence the ability to deliver the "metrics" ping as well.

To sum:
* We have a long list of things it isn't: 
    * Differences in what Glean vs Telemetry think a "channel" or a "version" is.
    * One weird client
    * Worse than it looks.
    * Preinit queue overflow.
* The best guess for what it could be (at-shutdown ping scheduling differences) doesn't quite pan out since there's just so many sessions that the recorded event should be sent at-startup.
* Mayyyybe since the sessions are so short (over half are 2min or shorter) Glean's not getting past preinit? BrowserGlue [says](https://searchfox.org/mozilla-central/rev/887d4b5da89a11920ed0fd96b7b7f066927a67db/browser/components/BrowserGlue.sys.mjs#2313) it guarantees init will run, though.

I'm at a loss. ...though, if all the sessions are short that Glean is never init until late enough in shutdown that it can't upload a ping... but no, it uploads "baseline" pings just fine.

I don't know why this is happening, and I can't think of a way to look at the data to get any more clues.

Back to Bug 1837230 Comment 4