OOM on Elasticsearch client

Hi All,

We are using ES 6.8.0 and we see that our applicaiton is going out of memory after certain crossing certain threshold of data processing. We see that there are lots of “io.netty.channel.ChannelOutboundBuffer$Entry” and many “elasticsearch[client][transport_worker]” threads. We tried tuning the underlying TransportClient with "transport.netty.worker_count” parameter but it seems to have no effect. Looks like once an threshold is crossed, the backlog starts building up on the client side leading to OOM.

Please let us know what other tuning parameters we can try both on server and client side or if there is any feedback mechanism is available so that we can have some control over our processing.

Thanks

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.