Lucas Kent
Lucas Kent
KafkaSinkCluster needs to follow the same rack aware model as CassandraSinkCluster as we want to use it for the same use case. ## Extended shotover nodes config The config needs...
We need a kafka sink that will allow us to run shotover on a different host from kafka. For this use case performance is a critical concern. ## message rewriting...
We are seeing some issues with kafka in a test setup that are not showing up in our integration tests. I've got a few ideas as to why it passes...
Currently we use rust-rdkafka as our kafka driver in our integration tests, however it does not support TLS. So if we want to integration test TLS support for the source...
`HashMap` is now a common idiom in transforms for keeping track of a generated response that needs to replace the response to a certain request. In order to optimize this...
blocked on: https://github.com/shotover/shotover-proxy/pull/1467 https://github.com/shotover/shotover-proxy/pull/1465 The first commit is performance neutral but for some reason the 2nd commit is heavily regressive:  From looking at a profiler we spend a lot...
In cassandra protocol v4 the max message is 256MB. At our expected batch sizes of 50 in transit requests, it would max at over 10GB which would OoM at a...
Running the same bench twice does not give stable results, neither locally nor on AWS But I'm focusing on AWS atm because that feels more important. network and disk IO...
Profiling is a bit of a mess right now. Every profiler collects and stores its results in a slightly different way. A possible refactor I'm thinking of is this: Build...
Cassandra routing in CassandraSinkCluster has shown up as a hotspot in profiling. `NodePool::get_replica_node_in_dc` shows up in the samply profiler as taking up 2.6% of execution time in the `cassandra,compression=none,connection_count=100,driver=scylla,operation=read_i64,protocol=v4,shotover=standard,topology=cluster3` windsock...