[08:45:42] <brouberol>	 I've submitted a patch stack to illustrate my point: https://gerrit.wikimedia.org/r/c/operations/deployment-charts/+/1084030/1
[08:56:02] <elukey>	 brouberol: o/ so basically you want a way to point the inbound traffic to the mesh envoy to select only a subset of the deployed pods
[08:56:54] <elukey>	 because it is a different kind of namespace, where not all pods are equal and able to reply to say HTTP traffic
[08:57:59] <elukey>	 I think it is a good idea, probably airflow will be the only use case for a while to benefit from this, but it seems to be a good compromise
[08:58:17] <elukey>	 ideally we'd segregate pods in namespaces in a different way
[08:58:30] <elukey>	 but I guess it would be painful for airflow
[09:01:25] <brouberol>	 that's exactly right yes. And indeed, it would be painful for airflow :/
[09:05:21] <elukey>	 okok I think that it is a good way to proceed, it is opt-in and you are just adding selectors
[09:05:52] <elukey>	 Janis in afk so maybe let's get an approval from Serviceops before proceeding, then I'd say we can do ti
[09:05:55] <elukey>	 *it
[09:06:01] <elukey>	 is it super urgent or can it wait?
[09:23:32] <brouberol>	 it can wait, in the meantime I've edited the services manually
[11:47:43] <elukey>	 brouberol: o/ I'd need to run the provision cookbook for dse-k8s-worker1009, lemme know when you have time to cordon/depool it (it will take max 10 mins but it will trigger a reboot)
[11:47:55] <elukey>	 no rush, even tomorrow
[13:17:02] <brouberol>	 Feel free to run it at any point
[14:06:06] <elukey>	 oook
[14:44:53] <elukey>	 brouberol: done!
[15:06:16] <brouberol>	 👍thanks!
[15:58:16] <elukey>	 cdanis: o/ aux-k8s-worker1003 runs on containerd now
[15:58:22] <elukey>	 (on bookworm)
[15:58:47] <elukey>	 it looks good, I'd say that we can keep this config for some days and then migrate the rest
[15:58:52] <cdanis>	 thanks!!
[15:58:57] <cdanis>	 ”the rest”
[15:59:19] <elukey>	 our cluster is big! :D
[16:05:06] <brouberol>	 niice