Vedran Lerenc

@shreyas-s-rao commented out-of-band on a Slack (I contacted him before I created this BLI proposal afterwards): > Isn’t HVPA being dropped for etcd? Yes, the plan is to generally drop...

Furthermore, isn't it the case that when and while the validation happens, ETCD itself is on hold (when the pod started, it hasn't even consumed any relevant resources yet). That...

Here are some sample stats on the 4 ETCD containers: ``` test_landscape_foo: etcd-main: etcd: metrics: allocatable: cpu: '8421' memory: 31.1T efficiency: used_requested_cpu: 27 % used_requested_memory: 39 % requests: allocatable_cpu: '674'...

FYI: I updated the ticket main description based on the above numbers. @shreyas-s-rao I hope we can get the first set of changes in quickly (the side cars only when...

It depends on what the chances are for the review @shreyas-s-rao. :-) It's not urgent and we'd like to bundle our changes to roll less often. If the review process...

Are there more concrete plans beyond that linked PR? Is there even capacity for this topic or are you now more involved with the Dashboard?

Same here, with the CPU patch (https://github.com/CompVis/stable-diffusion/pull/56).

We may need that feature not only because of **latency** and **cost**, but also for **functional/HA** reasons. Example: Zone 1: KAPI + KCM leader talks to KAPI in zone 3...

Gut feeling is similar as @unmarshall 's. If we don't have it, we pay with latency and cost at all times and the functional penalty now on top in case...

@dimityrmirchev Do you maybe have another link/docs where Google gives more details on GKE and this information? The page you referenced says: > This API is not yet intended for...