This bug is more about telemetry collection and establishing any decision logic on whether to load from network despite we have a (enough) fresh content in the cache when there is a good chance it would be faster when rather reload from the network (w/o revalidation). The simple start would be to have a new probe that shows a distribution between cache load time and network load time, both happening in parallel, broken by some basic set of preconditions we happen to know about the loading resource, like: priority class, estimated content type, size, ...? and maybe also: connection RTT, if known at the time of the request, connection type, ...?
After discussion in Vienna, we may have some more detailed plan. Let's use this as a tracking bug for all the particular tasks.
Assignee: nobody → honzab.moz
Summary: Experiment with racing network request against cached load → Tracking bug for network/cache racing
Sorry, wanted to assign to Michal (leader of this project)
Assignee: honzab.moz → michal.novotny
Whiteboard: [necko-active] → [necko-quantum][necko-active]
Bulk priority update: https://bugzilla.mozilla.org/show_bug.cgi?id=1399258
Priority: -- → P1
You need to log in before you can comment on or make changes to this bug.