Page MenuHomePhabricator

fragment-level cache hits take too long to get to AW
Open, Needs TriagePublicBUG REPORT

Description

Steps to replicate the issue (include links if applicable):

What happens?:
Each section takes about 20000ms to load 10 fragments, even when the fragments are all just a single space.

What should have happened instead?:
2000ms per fragment for a cache hit seems excessive.

From T422613:
"the existing fragment-level memcached caching in AbstractWikiRequest / CacheAbstractContentFragmentJob, which will continue to serve its original purpose of absorbing repeat fragment calls close to the orchestrator."

Even if this is protecting the orchestrator/evaluator from repeat calculations, the result does not seem to be getting back to AW quickly enough.

Event Timeline

Do you mean "takes to long to get to my browser"? I'm a bit confused by this as a request.

Sure, you can put it that way if you like. I expect the problem is true for everyone's browser though. Try loading the link above, you'll see how tortuous it is.