Jan Teske

I opened a PR here: https://github.com/MaterializeInc/materialize/pull/13247 I'm not sure it resolves the performance issue though. At least from my testing it doesn't. The effect is much less pronounced on my...

I can reproduce this reliably on my local machine. This is the test code: ```python c.up("materialized") c.wait_for_materialized() # 1. Create a remote cluster and verify that tests pass. c.up("computed_1") c.up("computed_2")...

@antiguru This looks like what was reported [here](https://materializeinc.slack.com/archives/C02PPB50ZHS/p1658489569228959). Can you try if this goes away with https://github.com/MaterializeInc/materialize/pull/13795? (Note: I don't actually know why this panic happens, just that reducing computed's...

I merged a fix for the issue where the since frontier of a storage collection was not correctly held back for slow replicas in a multi-replica setting. My hope is...

@aalexandrov I think you fixed similar overflows in the past. Can you take a look at this one?

This affects the test `dataflow_description_protobuf_roundtrip` as well ([for example](https://buildkite.com/materialize/tests/builds/41873#0183669b-e8e7-4d6e-becb-d55812672565)). I suspect both flakes have the same cause, namely proptest generating `DataflowDescription`s that are too big for the stack.

If we want to include introspection relations here, we need to introduce the unstable flag not only for tables, but also for sources, materialized views, and views (which is what...

I'd like to take the Porta cipher, if I may.

Another instance of this bug: https://materializeinc.slack.com/archives/C02PPB50ZHS/p1663600296010169

> Did you see this increase in idle load for debug or release builds? I did check that a while ago and didn't see a big increase in idle load...