abhishekrb19 opened a new issue, #19021:
URL: https://github.com/apache/druid/issues/19021

   ### Motivation
   
   The LoggingEmitter currently emits _all_ metrics it receives without 
filtering. On busy clusters with high query throughput, this can result in a 
very large volume of metric logs - on some of our clusters, we've seen ~3M 
metric logs in 15 minutes.
   
   ### Proposed changes
   - Add a file to capture the default metrics for the LoggingEmitter, similar 
to the Prometheus and Dropwizard emitters (e.g., 
`defaultMetricDimensions.json`).
   - This file can contain the list of allowed metrics.
   - In the future, we can extend this file to annotate additional user-defined 
dimensions, if needed.
   - In the LoggingEmitter `emit(Event event)` method, check the allowlist 
metrics map and decide whether to include the event or not.
   
   ### Rationale
   
   Typically, the LoggingEmitter is used in conjunction with a composite 
emitter that handles low- to medium-cardinality metrics, such as Prometheus, 
Dropwizard, or Kafka forwarding to a downstream system. Filtering out metrics 
that are already covered by these emitters would help reduce logging costs 
while continuing to rely on the appropriate emitter for low- to 
medium-cardinality metrics.
   
   ### Operational impact
   
   None. By default, all metrics will continue to be emitted for backwards 
compatibility with a runtime property. Operators can opt-in to enable filtering 
with the property.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to