Henning Jacobs

After trying suggestions in this thread (ffmpeg etc), I found that the "v4l2-sink" plugin works flawlessly (~0 latency) on Ubuntu 19.10 for my use case (Google Meet). Here the blog...

@markmandel I did not encounter this problem :thinking:

SGTM, a calendar integration would actually be nice :smile:

Yes, of course, go ahead! I would be very glad if you take care of this :smile: I'm not using Helm myself, that's why I refrain from taking over Helm...

Can you describe your use case? You want to have no downscaling by default, but in this case you can already just set the `downscaler/uptime` annotation and not set any...

@gregsidelinger that's not an issue with [pykube-ng](https://github.com/hjacobs/pykube) which I maintain. I also generate events in [kube-janitor](https://github.com/hjacobs/kube-janitor).

@propalparolnapervom thanks for the feedback, this is also related to #119.

Yes, this behavior is expected, but we can obviously discuss what makes sense in general.

This was implemented and documented on https://github.com/hjacobs/kube-resource-report#custom-links TODO: also allow linking nodes (e.g. to a monitoring dashboard), ingress and other entities

Good point, the main driver is adapting CPU/memory requests to match actual usage (requests - usage * cost_per_unit is the formula to get "slack costs").