TheCodeAssassin

One thing worth mentioning is that reformatting the node with kernel parameters to join a different omni account works flawlessly: https://tca0.nl/NGs

@Unix4ever ``` metadata: namespace: default type: MachineStatuses.omni.sidero.dev id: 30313939-6432-3866-3661-373136306465 version: 12 owner: MachineStatusController phase: running created: 2024-03-19T16:09:02Z updated: 2024-03-19T20:15:16Z labels: omni.sidero.dev/arch: arm64 omni.sidero.dev/available: omni.sidero.dev/connected: omni.sidero.dev/net: 1Gbps omni.sidero.dev/platform: metal omni.sidero.dev/reporting-events: omni.sidero.dev/storage:...

> When using Meilisearch on Kubernetes, the instance will sometimes take the host machine RAM as the reference and not the limits set by the container. I suggest using the...

@ManyTheFish that didn't solve the problem, it still crashed today. Now it did die with code 137: ``` Last State: Terminated │ │ Reason: Error │ │ Exit Code: 137...

> Hello @thecodeassassin, I am stunned by this error. The only way Meilisearch crashes without returning an error is by being killed. Most of the time, it is because of...

> Hello @thecodeassassin, What are the `periodSeconds` and the `failureThreashold` values? I suggest putting both of them to 5. > > Then, the issue could come from other pods on...

> Hello, > > would setting the env var `MEILI_EXPERIMENTAL_SEARCH_QUEUE_SIZE=0` mitigate the issue and reduce the number of crashes? We are already running it with this setting: ``` spec: containers:...

> alright, thank you. > > > > are you using the vector search? Hi, No we are not using vector search.

One thing worth adding to this is that the amount of restarts do decrease if the number of searches decrease. We have implemented a caching proxy in front of Meilisearch...

> I am unsure of whether I should open a new issue or post here, so please tell me if this isn't the place. > > We're still experiencing this...