I have a bit of a problem. I have a typesense cluster behind a gcp load balancer as an external network endpoint group. I upgraded our version to the latest 0.25rc, but that must have broken something because all requests timed out. Now the cluster version is reverted to 0.24.2 (even though that wasn't an option in the version dropdown). Is it possible that something about 0.25rc doesn't work with the way we've setup networking/proxying?

Hey Walter, I just hit send on an email! To add more context, 0.25.x.rc has some changes which makes it not downgrade-able to previous versions of 0.24, except a special version 0.24.2 that explicitly allows for downgrades. So we’ve set your cluster to this version instead.

as always you're two steps ahead. Do you think something about the 0.25rc version could be interfering with how we proxy?

I upgraded our version to the latest 0.25rc, but that must have broken something because all requests timed out.

Hmmm, that sounds strange… especially with

0.25.0.rc41

, since we have a few users already running this build

ok thank you that at least provides an avenue for debugging.

Let me know if you can replicate the issue on another node, and we can take a closer look if there’s an issue on the Typesense side

wait but after downgrading wouldn't that be the same issue?

I would imagine so - I was just about to ask - does the issue not exist now after downgrade?

no it works again now

Not at the moment… We don’t want to publish

0.24.2

publicly just yet, before we release 0.25.0 GA… so someone on the Typesense side would have to downgrade you to this version until then

ok cool. What i'll do then is replicate the proxying in our dev environment instead. I was trying this on prod where I don't proxy typesense through our domain.