I think this is probably related to #8 .
I was tracking down a bug in our plan and we were awaiting a series of client.record_metric() calls. There were quite a few of these, lets say like 50. It was taking 4 seconds and holding up the executor.
I assume the culprit is influx db. Currently I'm spawning this collection of calls in their own task, but I would expect writing to a db for <100 queries should take less than 4 seconds.
Raising for awareness.
I think this is probably related to #8 .
I was tracking down a bug in our plan and we were awaiting a series of
client.record_metric()calls. There were quite a few of these, lets say like 50. It was taking 4 seconds and holding up the executor.I assume the culprit is influx db. Currently I'm spawning this collection of calls in their own task, but I would expect writing to a db for <100 queries should take less than 4 seconds.
Raising for awareness.