db_timeout needs to be configurable per pageid

Bug #609564 reported by Robert Collins on 2010-07-24
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Launchpad itself
Robert Collins

Bug Description

We want to prevent backsliding on pages that have had their performance sorted out, prevent resource hogging by malicious clients or users, and iteratively lower the timeout down to 5 seconds.

However, some pages will always be significantly more expensive than others, and until we have good ways of doing that work out-of-appserver-transaction, the db timeout mechanism will block all progress on lowering the timeout, because those pages will break if we do.

I'd really like it if one could list page ids in the config file, so that the sysadmins can raise the limit for selected pages, rather than all pages, when there is an issue. This isn't ideal, and I realise its both more work and impacts on report generation and analysis.

However, the alternative is to block lowering the global timeout on a handful of problematic pages, and that doesn't make sense to me. I'm filing this as high, because of the blocking nature of it. However - I'm also very likely to write the patch for this myself in the next few days, sleep & travel permitting, so I'm going to assign it to me as well. If someone has time and wants to do it, please just reassign it to yourself.

Robert Collins (lifeless) wrote :

pageid: scope permits this.

Changed in launchpad-foundations:
status: Triaged → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers