Lucas Kent
Lucas Kent
kafka benchmarks are currently quite wonky. We currently use the cpp rdkafka driver which does not map well to rust async. Things to do from here: * add benches using...
This gives more stable results when no OPS cap is set, but when I cap the OPS the results jump significantly above the cap which is very wonky. Leaving as...
Currently we have our metrics split up like this: * `shotover_transform_total_count` * `shotover_transform_failures_count` where failures_count is a subset of total_count. I think it would be more intuitive if had them...
It should fail with a message like: ``` KafkaSinkSingle: the destination_port is the same as the port used by `KafkaSource`. These ports must be unique as shotover and the destination...
I was reading a blog post during my holiday and realized that we set tcp nodelay for clientshotover connections but we dont set it for shotover DB connections. Should we?
I've observed that shotover will shutdown without an exit code (causing TokioBinProcess to fail the test) when we startup shotover and shut it down immediately after. I demonstrated this was...
Analysis of profiling results from `cargo windsock --profilers samply --name kafka,shotover=message-parsed,size=100KB,topology=single` 
WIP. There are two places shotover would benefit from cross-instance communication: * Improved load balancing * Simplified horizontal scaling This issue addresses only "Simplified horizontal scaling" but our design should...
progress towards: https://github.com/shotover/shotover-proxy/issues/868 waiting on https://github.com/shotover/docker-compose-runner/pull/2 All though, since even with that PR its having troubles, maybe its best to abandon that approach and just drop the `DockerCompose` and recreate...
Intermittent test failures are now occurring pretty commonly, time to do some investigation and flush them out. ## cassandra_int_tests::cluster_single_rack_v4::case_2_datastax Oh oops, I wasnt thinking and overwrote this with a manually...