I have 500 or so cases of the telemetry event replication:medic:to:failure:reason:error.
Is there a path to understanding the error which caused this in more details?
Hi @kenn
You won’t be able to get additional data through telemetry, which is just a counter of these events happening. There is no additional granularity for the error type (though we do have separate categories for fetch and response parse failures).
Relevant webapp code: https://github.com/medic/cht-core/blob/496a2bc9d945ed8cb0bb9e503a16d777fbecc465/webapp/src/ts/services/db-sync.service.ts#L430
I suggest you check feedback docs to get more details about these failures.
I would be very interested in hearing your findings.
Here and Here it looks like telemetry events are recorded without creating a feedback doc. Am I reading that right? Or should I expect one feedback doc for each replication failure with reason:error?
Recording telemetry and registering a feedback doc should be independent. Feedback docs get created automatically on uncaught errors.
Are you not finding feedback docs for those replication failures? Please share your results.
I’ve created an issue yesterday after noticing that failed report submissions don’t create feedback docs: Caught errors do not generate feedback documents. · Issue #8089 · medic/cht-core · GitHub .
We should make sure feedback docs get created when unexpected errors happened, even if we got better at catching them.