Major Description
While running system tests locally, I've noticed excessive logging of the Telemtry Reporter. This I believe was introduced in KIP-714.
[2024-01-15 09:44:16,911] DEBUG For telemetry state SUBSCRIPTION_NEEDED, returning the value 224678 ms; the client will wait before submitting the next GetTelemetrySubscriptions network API request (org.apache.kafka.common.telemetry.internals.ClientTelemetryReporter)
This is logged several times per ms - Also, given the amount of log being emitted, can we also check the CPU profile to see if there's a process running a tight loop?
Update
-----------
Looking from the beginning, is this caused by the following?
DEBUG The broker generated an error for the get telemetry network API request (org.apache.kafka.common.telemetry.internals.ClientTelemetryReporter)
146 org.apache.kafka.common.errors.UnsupportedVersionException: The node does not support GET_TELEMETRY_SUBSCRIPTIONS
Attachments
Issue Links
- links to