Embedders can't be sure they've run long enough to process some pending pings
Categories
(Data Platform and Tools :: Glean: SDK, enhancement, P4)
Tracking
(Not tracked)
People
(Reporter: chutten, Unassigned)
References
(Blocks 3 open bugs)
Details
Most of our embedding applications usually run for long enough that the SDK can process and attempt the upload of at least a few of its pending pings... averaged over a suitable interval, at least.
But that's not all of them. Background tasks in Firefox Desktop can spin up, record metrics, submit pings, and spin down so fast that our heads'd spin... and their pings might never catch up.
Knowing how to handle startup and shutdown is a very embedding-level problem. (e.g. It's FOG's problem, not the SDK's) But in order for the embedder to handle shutdown, it needs to know the SDK's had a chance to process its pending pings first.
This ping is about finding and building and testing and documenting a way to communicate to the embedding layer that the SDK's had a chance to do this. Maybe through a special uploader status?
Updated•2 years ago
|
Updated•2 years ago
|
Description
•