@wyri When we started building ARM packages, and then container images, I was shocked at how many little things might not work - some packages did not have ARM versions, etc. Amazing how much you take for granted always developing for AMD64.
My efforts to put all my #GitHub Action Workflows into a centralized repository and make where to run configurable on the ones for private projects is paying off. Most things can now run on my home cluster, but scaling is still somewhat slow:
Been going through my photos to find the photos I need to for the 3 blog posts I wrote about it yesterday. Came across some random photos, starting with the first node getting set up, with GUI for some reason:
Wrote 3 blog posts today 🫣 . Finishing off the 4th, and doing GitHub sponsors update after. Will be previewing two of those blog posts in there as one is a bit short: https://github.com/sponsors/WyriHaximus
Loving #PhilipsHue's long press to turn all the lights off. With some node-red and #HomeAssistant, I just extended that to our #lego light kits and display lights.
It was slightly infuriating to do because I had to figure out which event it was and to do that I had to look at all events coming from HA just to find "hue_event" among all "state_changed" events. It was simple after that to find the "long_release" event type for the "hue_event".
@benoit_badrignans@wyri Not sure if it applies, but I do think so, but have you ever looked at the events monitor in the development tools?
I never used Node-red but this should be relatively easy to do using just the HomeAssistant automations and depending on your Zigbee gateway it might even be a build in trigger the you can select from a drop down menu.
@secupriv@benoit_badrignans Yeah it applies, thanks both found them in both locations. Will go there first next time because the "hue_event" is listed. (Not that in matters much but not using Zigbee directly but the official Hue bridge.)
Using node-red in such a way that I get all events, filter all that are of the long_release type, and then trigger the service calls.
My plan for this afternoon was to finish my centralized reusable #GitHub Actions workflows for releasing projects. But some outage is keeping me waiting 😅. (Can even see the self hosted runners spin up but not getting any jobs.)
Having to recycle your #kubernetes nodes because the #HPA isn't scaling up your #Mastodon#sidekiq so you're now lagging 15K jobs behind looks like this in the #UI:
Ok ok I admit that I recycled my nodes because I could. Kicking the pod probably also would have fixed it but they where up for 82 days and it felt fun