Hi Robert, Thank you very much for moving this forward!
On 12-05-29 12:43 AM, Robert Collins wrote: > So, I finally made time today to progress the 5 second timeout effort > a bit more today. > > As of now, the *default* hard timeout for LP pages is 5 seconds. Pages > with a 99th percentile larger than that have had a higher timeout set: > 9 seconds, if the page's 99th percentile render time was over 9 seconds > Or the nearest second higher than the pages 99th percentile render time. > > Specifically, a pageid with a 99th percentile render time of 6.1 > seconds would end up with a 7s hard timeout. > > No change in developer behaviour is needed - this just enforces the > hard timeout portion of the previously agreed upon decision: that new > pages would be held to our target performance of 1s for their 99th > percentile render time, with a hard timeout of 5 seconds. > > The way this was done, was to record the page ids for all the slow > page ids; the DB request to retrieve the 100 odd extra rules takes > 3ms, and the render time for a basic page was not impaired on > qastaging, so I don't expect any fallout. This is also good news for > folk with lots of feature flags / rules: the system is scalable well > beyond how we are using it today. > > Cheers, > Rob > > _______________________________________________ > Mailing list: https://launchpad.net/~launchpad-dev > Post to : launchpad-dev@lists.launchpad.net > Unsubscribe : https://launchpad.net/~launchpad-dev > More help : https://help.launchpad.net/ListHelp -- Francis J. Lacoste francis.laco...@contre.com
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Mailing list: https://launchpad.net/~launchpad-dev Post to : launchpad-dev@lists.launchpad.net Unsubscribe : https://launchpad.net/~launchpad-dev More help : https://help.launchpad.net/ListHelp
