Seeing repeated dependency calls to Cosmos showing up as "failed" in app insights with a result code of 0 and an error of "The operation was aborted".

Seems to happen on POST requests.

Doesn't seem to trigger any exceptions from the SDK, so presumably this is something that's happening internally.

What causes this?

Is it something as basic as determining whether or not to use a query plan? Something else?

Looking to cut down on noise in logs:

• If there anything on our side that can cut down on the # of calls?
• If not, is there a combination of criteria that we should be filtering on in our telemetry processor to flag these as success? (presumably there could be other operation aborted errors within the SDK that might be actual errors)