Open Bug 1639586 Opened 5 years ago Updated 4 years ago

Consider limiting the number of outstanding timer ids in Timing Distributions

Categories

(Data Platform and Tools :: Glean: SDK, defect, P4)

defect

Tracking

(Not tracked)

People

(Reporter: chutten, Unassigned)

Details

(Whiteboard: [telemetry:glean-rs:backlog])

Every time we start a sample for a timing distribution we allocate memory. Maybe we shouldn't do this unboundedly because code that leaks timer ids (by starting and never cancelling or stopping them) might not be caught by the compiler (unlike in Rust!) or user instrumentation tests in all our language bindings.

Maybe we should pick a number (32? 256? 1024? I like powers of two) and error out the first time/each time we hit that limit?

Priority: -- → P3
Whiteboard: [telemetry:glean-rs:m?]
Whiteboard: [telemetry:glean-rs:m?] → [telemetry:glean-rs:backlog]
Priority: P3 → P4
You need to log in before you can comment on or make changes to this bug.