Setup some alert mechanism when some 'critical' cron jobs fail
Open, NormalPublic


I noticed that on T187053: Setup puppet cron to delete old data daily the script we use to purge old afl_ip data from the abuse log to comply with the WMF privacy policy and the data retention policy broke in 2016 and it was not until today that we noticed that when we were working on making some of that data avalaible per T160357: Allow those with CheckUser right to access AbuseLog private information on WMF projects. This means WMF has been inadvertently collecting private data for more than 90 days. To avoid that in the future, please configure the purge_abusefilter and the purge_checkuser mediawiki cron jobs to email someone to let them know when the script has crashed, not run for some reason, etc. Please also expand this practice to any other maintenance script which could be tagged as 'critical' by its function. Thanks.

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptFeb 12 2018, 6:17 PM
fgiunchedi triaged this task as Normal priority.Feb 13 2018, 11:23 AM
fgiunchedi added a subscriber: fgiunchedi.

I agree with the general sentiment/idea, though implementation wise we should be going through icinga alerts or phab tasks and not emails.

icinga alerts or phab tasks works for me as well