1048640 - Alert memshrink-telemetry-alerts list about telemetry changes relating to memory usage

Status: RESOLVED FIXED

(Toolkit :: Telemetry, defect)

Description

[Nicholas Nethercote [inactive]]

I'd like to be informed when various memory-related telemetry histograms
change.

Comment 1

[Nicholas Nethercote [inactive]]

Attachment: Alert nnethercote about telemetry changes relating to memory usage

This seems reasonable, but I haven't done any kind of testing because I don't
know how to do so.

Comment 2

[Vladan Djeric (:vladan)]

Comment on attachment 8467459 [details] [diff] [review]
Alert nnethercote about telemetry changes relating to memory usage

Would it be better to create an email alias for these alerts so others on MemShrink could get them as well?

Comment 3

[Nicholas Nethercote [inactive]]

Hmm, probably. How should I go about that?

Comment 4

[Vladan Djeric (:vladan)]

It's actually pretty easy with ServiceNow.

1. Log in to https://mozilla.service-now.com/navpage.do
2. "Updating something" -> "Mailing Lists"
3. "Create a new mailing list" -> "Type: Distribution List"

Comment 5

[Nicholas Nethercote [inactive]]

Attachment: Alert memshrink-telemetry-alerts list about telemetry changes relating to memory usage

I've requested the creation of a list and updated the patch accordingly.

Comment 6

[Nicholas Nethercote [inactive]]

Comment on attachment 8469844 [details] [diff] [review]
Alert memshrink-telemetry-alerts list about telemetry changes relating to memory usage

Review of attachment 8469844 [details] [diff] [review]:
-----------------------------------------------------------------

Carrying over r+.

Testing these changes is difficult. What would happen if somebody typo'd an email address? They'd never get alerts and probably wouldn't realize...

Comment 7

[Nicholas Nethercote [inactive]]

I'll land this once I've received confirmation that the list has been created.

Comment 8

[Nicholas Nethercote [inactive]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/50ac8d9ddf27

Comment 9

[Vladan Djeric (:vladan)]

(In reply to Nicholas Nethercote [:njn] from comment #6)
> Testing these changes is difficult. What would happen if somebody typo'd an
> email address? They'd never get alerts and probably wouldn't realize...

Actually we've already had an email address typo happen in bug 1047568. Luckily, we get alerts from Telemetry infrastructure whenever something goes wrong (mail server throws an error, submission count drops, etc) so it was caught fairly quickly. Of course, you could also just send a message to the email address you're entering in Histograms.json and verify the right people have received it.

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/50ac8d9ddf27

Comment 11

[Nicholas Nethercote [inactive]]

> Luckily, we get alerts from Telemetry infrastructure whenever something goes
> wrong (mail server throws an error, submission count drops, etc) so it was
> caught fairly quickly.

Good to know!

> Of course, you could also just send a message to the
> email address you're entering in Histograms.json and verify the right people
> have received it.

Yep, I did that. So that tests that the email address is correct, but there are other things that could go wrong. E.g. something could go wrong with the machine that detects the regressions and it fails to send the emails. But maybe the checking you mentioned above catches these cases.