Alexander Kuzmenkov

I'd prefer `Ensure` because all-caps is inconvenient to type.

Everything integrated in other PRs.

> Looking through the remote_copy, I see that we use a replication factor of 2 (good), but it might be a good idea to test with a few other replication...

Turns out the text COPY passthrough is just totally broken :weary: https://github.com/timescale/timescaledb/issues/4761

> Is there an agreed way forward on this issue even if addressing it isn't imminent? A related timescaledb-toolkit issue was closed last year pending work on 'multi-value timeseries' but...

We merged some distributed COPY improvements into 2.9 which is going to be our next release. This issue is about INSERT, so keeping it open for now.

> We should test and document any limitations around using < > operators. > > How about a test with CASE expression? You mean, the fact that they are not...

@antekresic can you explain more about this issue? What row was it trying to lock that was invisible?

It works if you qualify the call with schema name in `test_call()`, i.e. `return public.test_return_text('test');`. AFAIK in the continuous aggregate functions, the search path is locked down to builtin and...

With LIMIT, the slower paths should normally get disabled through the startup cost, right? Maybe we can tweak the costs there so that it works? Our cost model for DecompressChunk...