Alex Howle

> The configurable timeout is only there to control how long it'll hang out looking for a prompt to respond to before rolling to the next loop iteration (and allowing...

I'm having to spin-up on `pexpect` and the way stdout is hooked to get the CLI output so I may be off the mark, here - would there be any...

I've got a repro for this issue - no container required. A combination of the number of hosts in the inventory (some of which are unreachable), and the time spent...

We've already moved the event handler logic to another thread and the problem has gone. Nothing like a slow-consumer problem to remind you why you need to be quick on...

It's like the observation of burrow changes the results.

I've performed more testing and it seems the gRPC disconnection is irrelevant; the aged StartTimeUnixNano is used regardless.

> @alexchowle care to give #3978 a run? I've ran the reproduction scenario and am still seeing the same issue. ``` Observation # 1 ----------------- Metric #0 Descriptor: -> Name:...

I only noticed this issue because my vendor backend uses `StartTimeUnixNano` as the observation timestamp. My service has a very low throughput.

> @alexchowle I have created a branch that I think should fix your issue: #4002 > > I am just adding a [commit](https://github.com/open-telemetry/opentelemetry-python/pull/4002/commits/c424be3ae0390f384aa9f41d2870c6ac01ad69e5) so that the previous collection time gets...

Maybe a result of #1343 ?