Matt Lord

@anshikavashistha are you still working on this?

I'm not sure what this means and if it's needed. We should clarify some before spending time on it.

cc @vitessio/query-serving

I don't have anything against this, but can you list out what you see as the practical advantages this offers @morgo? Moving the binary flags to config files is nice...

Thanks, @chrisplim ! ❤️

I would note that this kind of topo work is often much more challenging than it would seem at first glance. I would recommend you have a PoC (proof of...

> On clusters with 1000s of tablets this becomes a scalability problem for the topology store that has to handle topo Get calls for all of the tablets fetched. I...

@rvrangel I don't see where that happens here: https://github.com/vitessio/vitess/blob/da4b81d1bab0864c7ddee2eb58c872e251c30f5a/go/vt/discovery/topology_watcher.go#L148-L203 Is there a test which demonstrates the problem we're trying to solve here? We could certainly do the cell based filtering...

My point is that the premise here seems somewhat flawed AFAICT. The premise seems to be that we get every tablet record *individually*, but I don't see that we do...

Hi @aep ! > it isnt actually missing. forcePathStyle is just missing from the docs here https://github.com/planetscale/vitess-operator/blob/main/docs/api.md Ah, I was going to say that I knew some people were using...