Opened 7 years ago

Last modified 4 years ago

#2032 new enhancement

Report tracebacks as singular metrics events

Reported by: bdbaddog Owned by:
Priority: patches-accepted Milestone: 0.9.+
Version: 0.8.4p1 Keywords: metrics
Cc:

Description (last modified by sa2ajj)

It would be useful to log tracebacks in a separate file. Rather than having to grep for them from the twistd.log's

So perhaps: traceback.log or exceptions.log

Change History (8)

comment:1 Changed 7 years ago by dustin

  • Milestone changed from undecided to 0.8.+
  • Summary changed from Log tracebacks in seperate logfile (not in twistd.log) to Log tracebacks in separate logfile (not in twistd.log)

This is a good idea, but I probably won't work on it myself. Anyone else?

comment:2 Changed 6 years ago by dustin

  • Keywords metrics added
  • Priority changed from minor to patches-accepted

This would be cool to handle via metrics alerts.

comment:3 Changed 4 years ago by dustin

  • Milestone changed from 0.8.+ to 0.9.+

Ticket retargeted after milestone closed

comment:4 Changed 4 years ago by sa2ajj

  • Description modified (diff)

Dustin, I do not quite understand how this ticket is metrics related. Could you please elaborate?

comment:5 Changed 4 years ago by dustin

Metrics can handle singular events, as well as time-series data. I was thinking of making the exceptions into singular events.

comment:6 Changed 4 years ago by sa2ajj

  • Summary changed from Log tracebacks in separate logfile (not in twistd.log) to Report tracebacks as singular metrics events

Then the title is rather misleading: it's not about logging tracebacks in a separate file, but about make the exceptions into singular events :)

(I'll update the title then.)

And I think it should be in addition to the current way so the tracebacks are still logged in twistd.log file.

comment:7 follow-up: Changed 4 years ago by dustin

True, I was just suggesting a possible way to solve this. Simply logging the tracebacks to a separate file would probably make bdbaddog happy though :)

comment:8 in reply to: ↑ 7 Changed 4 years ago by bdbaddog

Replying to dustin:

True, I was just suggesting a possible way to solve this. Simply logging the tracebacks to a separate file would probably make bdbaddog happy though :)

Yes. The main goal would be a way to automate watching for any exceptions. For one client this has been caused by bad logic in an email formatting function, which meant that failure notifications weren't getting sent out and thus missed.

Any solution which would allow the following would work:

  • detecting when a new exception has happened
  • extracting each exception individually in some way which wasn't too complicated.

Thanks!

Note: See TracTickets for help on using tickets.