[QA] an experiment: fatal error notices on beta labs sent to this mail list

[Nikolas Everett]

On Mon, Jan 27, 2014 at 10:30 AM, Chris McMahon <cmcmahon at wikimedia.org>wrote:

>
>
>
> Hello QAers,
>
>
> We are about to try an experiment.  We have a small script that checks for
> fatal errors on our test environment at beta labs.  We've set up that
> script to send notifications to this list.
>
> I am concerned that it might be too much noise on a list that generally
> has a lot of signal.  If you find yourself annoyed by the messages, please
> speak up!  We can alter the frequency, or even change the email recipients.
>  I'd rather move the fatal notices than have people unsubscribe, so do
> complain if you feel the need.
>

Can you squash them together like they do in fatalmonitor?  For reference,
this is the source of fatalmonitor:
watch "tail -n 1000 /home/wikipedia/syslog/apache.log | grep
'PHP\|Segmentation fault' | grep -v 'filemtime\|failed to mkdir\|GC cache
entry\|cache slam averted\|SHA-1 metadata' | sed -r 's/\[notice\] child pid
[0-9]+ exit signal //g' | sed 's/, referer.*$//g' | cut -d ' ' -f 7- | sort
| uniq -c | sort -rn"

With that I can tell at a glance if the fatal is "mine".

Nik
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wikimedia.org/pipermail/qa/attachments/20140127/b6b3ef58/attachment.html>