Michael Harris

@mpraglowski Yes that should work for me!

@mpraglowski Is it okay if we move it back by 30 minutes? I have to bring a pet to the veterinarian in the morning and I will likely not be...

I'm running into what looks to be the same issue. We have Meilisearch 1.15 deployed on a k8s cluster and it only has a dozen documents in it. We're still...

Unfortunately it's _much_ easier to reproduce in a container environment because the hard limits immediately kill it. I haven't tried a standalone environment. In my case it doesn't require a...

I apologize for the delay, it should be harmless to share them publicly if I change a couple of minor details. I'll try and pull them for you today. Edit...

@taieb-tk that might be related. I noticed the issue when re-indexing documents. The flow was something like this: 1. Initially added the documents with only the primary key configured 2....

Seems like it's maybe no longer relevant since you made a bunch of progress, but to answer your question anyway (thank you for looking!) the fields I had filterable at...

Your finding looks plausible, that's a common issue in container environments and the host we're on is part of a larger cluster so it would be easy to see a...

I also don't see any crate that provides a unified abstraction for this. It does look like the sysinfo crate being used supports cgroups as well though: https://docs.rs/sysinfo/latest/sysinfo/struct.System.html#method.cgroup_limits It'll return...

> We then had to hotfix an increase to the limit. Yeah we deployed a change to increase the limit. We were just trying to use other methods first since...