By using AWS re:Post, you agree to the Terms of Use

Throttling a Client/Index from OpenSearch (ElasticSearch)


I have the following setup for logs:

  • Fargate Service -> Firelens/Fluent Bit -> OpenSearch (ElasticSearch) Cluster or
  • Fargate Service -> Cloudwatch -> OpenSearch (ElasticSearch) Cluster

I know about Fluent Bit Throttle, but that is on client side. What would be good mechanism on OpenSearch (ElasticSearch) Cluster side to throttle that Client/Index if he pushes to much data?

asked 9 months ago105 views
1 Answer
Accepted Answer

OpenSearch throttles by default when the respective threadpool queues are full. High threadpool queues indicate the that the incoming request rate is greater than what the cluster is provisioned to handle.

If you're going beyond this, you would have to setup a feedback mechanism by considering metrics such as CPUUtilization, JVMMemoryPressure, etc. You can set a CloudWatch alarm over the metrics to notify you via SNS as well as configure SNS to invoke a Lambda function to reduce the rate on the client side as required.

answered 7 months ago
  • The client might be in another account. The Lambda should execute there and reduce the rate there or should is there a way to to set some rule in OpenSearch to limit and incoming DataStream?

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions