Michael Penick

@jshook In practice, GOAWAY is common (in fact many HTTP/2.0 servers send this every N requests for a persistent connection see http://nginx.org/en/docs/http/ngx_http_v2_module.html#http2_max_requests and http://nginx.org/en/docs/http/ngx_http_core_module.html#keepalive_requests). The problem here is that `java.net.http.HttpClient`...

This https://github.com/datastax/cql-proxy/pull/36 completes part of this issue, but we still need to pull in the k8s guide.

It lives in the proxy. It's currently infinite on both accounts. This sounds terrible, but in practice it's not as bad as it sounds because it tends to be *really*...

Sized capped at ~100MB: https://github.com/datastax/cql-proxy/pull/28/commits/147ef7239c791a6abd6511443c58be413a3484f0

The lib the proxy was using didn't support this, but has since been fixed: https://github.com/datastax/go-cassandra-native-protocol/pull/26 I'll upgrade and push a new release.

Thanks for the report. I'll work on reproducing the issue tomorrow.

@datalatics-official Could you please describe your workload? Are you sending multiple requests asynchronously? Also, those numbers for connections per host are really high. The driver does coalescing of requests and...

I'm unable to reproduce using driver v2.1.10.3 using the following for load: ```java package org.example; import com.codahale.metrics.ConsoleReporter; import com.codahale.metrics.MetricRegistry; import com.codahale.metrics.Timer; import com.codahale.metrics.Timer.Context; import com.datastax.driver.core.Cluster; import com.datastax.driver.core.ConsistencyLevel; import com.datastax.driver.core.HostDistance; import...

Thanks for trying it out. Couple questions: * What version of the proxy? (`0.0.4`?) * Are you using it against Astra or Cassandra? * ".25 million reads per hour" is...

> .25 million reads per hour `1e6*.25/3600 == ~69.4 request/s` Is that right? I've run much bigger workloads than this w/o issue. Maybe the request/response payloads are much bigger?