The calculation of 'hangs_normalized' in the daily crashes cron is not being used. It was an attempt to account for throttling of the hang crash pairs to get a real count of the number of hangs. It doesn't do the job in a statistically correct way. It's use is deprecated and the code ought to be removed from the cron. strike this sql: UNION -- for HANGS_NORMALIZED SELECT count(subr.hangid) as count, %s, subr.prod_id, subr.os_short_name, timestamp without time zone %s FROM ( SELECT distinct hangid, p.id AS prod_id, substring(r.os_name, 1, 3) AS os_short_name FROM product_visibility cfg JOIN productdims p on cfg.productdims_id = p.id JOIN reports r on p.product = r.product AND p.version = r.version WHERE NOT cfg.ignore AND timestamp without time zone %s - interval %s <= r.date_processed AND r.date_processed < timestamp without time zone %s + (interval '24 hours' - interval %s) AND cfg.start_date <= r.date_processed AND r.date_processed <= cfg.end_date AND hangid IS NOT NULL ) AS subr GROUP BY subr.prod_id, subr.os_short_name strike: the corresponding line in the 'parameters' data tuple, too
Component: Socorro → General
Product: Webtools → Socorro
Looks like this was fixed a while ago, I don't see this now.
Status: NEW → RESOLVED
Last Resolved: 5 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.