Closed
Bug 1281765
Opened 8 years ago
Closed 3 years ago
Send reminder emails for expired probes to alert_emails recipients too
Categories
(Data Platform and Tools :: Monitoring & Alerting, defect, P3)
Data Platform and Tools
Monitoring & Alerting
Tracking
(Not tracked)
RESOLVED
DUPLICATE
of bug 1610616
People
(Reporter: gfritzsche, Unassigned)
Details
Currently there are apparently many expired probes and accompanying code which don't get removed. We could send regular reminder emails about them so they are not forgotten about. Bug 1261400 should also help here once the bug triaging process is fully in place.
Comment 1•8 years ago
|
||
We send regular reminder emails but only to telemetry-alerts, not to the alerts_emails fields specifically. [1] Not sure why (before my time) but we should figure out why we made the decision to not send it to alerts_emails and, if that reason is no longer valid, send them to alert_emails.
Comment 2•8 years ago
|
||
[1]: https://github.com/mozilla/cerberus/blob/master/alert/expiring.py#L127
Reporter | ||
Updated•8 years ago
|
Summary: Send regular reminder emails about expired probes → Send reminder emails for expired probes to alert_emails recipients too
Comment 3•8 years ago
|
||
to discuss with DDurst.
Comment 5•8 years ago
|
||
I assume the reason is because either alerts_emails isn't required? Anyway, it makes sense to me that if we have addresses there, we should send to them as well. I also assume it's not the case that everything has to be removed -- so this is just about ensuring that nothing falls through the cracks. RyanVM may have a stronger opinion on this than I do.
Flags: needinfo?(ddurst)
Comment 7•8 years ago
|
||
This sounds like a good idea to me, albeit with limited knowledge of the history that led to the current state of affairs. My understanding is that this is one of the issues that data review was intended to fix going forward (having a clear idea of ownership of a probe throughout its lifecycle). It's unfortunate that we have some probes without alert_emails set, but maybe we can handle those differently as-needed (hg blame should be able to tell us who needs bugging in any case). Anyway, this sounds good to me. We don't want to be leaving a bunch of cruft from old probes laying around IMO.
Flags: needinfo?(ryanvm)
Reporter | ||
Comment 8•8 years ago
|
||
alerts_emails is required for all new probes (we enforce it at build time and via data collection review). Only old probes from before we added that field have exceptions for that rule. Identifying emails/owners for all of these would be a time consuming task, which is why we still have them without.
Comment 9•8 years ago
|
||
to Chris, setting p2, feel free to change as needed.
Assignee: nobody → chutten
Priority: -- → P2
Updated•7 years ago
|
Component: Metrics: Pipeline → Monitoring & Alerting
Product: Cloud Services → Data Platform and Tools
Reporter | ||
Updated•7 years ago
|
Assignee: chutten → nobody
Priority: P2 → P3
Comment 10•3 years ago
|
||
chutten - do you know if part of the "getting to glean -FoG plan" includes removing dead probes from the code?
If yes we can just close this out - and add as part of the "migration plan" docs for teams as they migrate.
Flags: needinfo?(chutten)
Comment 11•3 years ago
|
||
Oh goodness, this is something the Probe Info Service already does, and has done for a couple of years. Guess this bug got left behind.
Closest to a duplicate bug I can find is bug 1610616, so let's use it.
Status: NEW → RESOLVED
Closed: 3 years ago
Flags: needinfo?(chutten)
Resolution: --- → DUPLICATE
You need to log in
before you can comment on or make changes to this bug.
Description
•