A long exposure of the rocky tip of Peveril Point in Swanage, Dorset.
#photography #dorset #peverilpoint #swanage #visitdorset #uk #water #coast #rocks
http://dailyphoto.creativesplurges.com/2025/05/21/peveril-point-5/
experimental data visualisation and animation. more @ https://www.beyondwatts.com
I’m searchable
A long exposure of the rocky tip of Peveril Point in Swanage, Dorset.
#photography #dorset #peverilpoint #swanage #visitdorset #uk #water #coast #rocks
http://dailyphoto.creativesplurges.com/2025/05/21/peveril-point-5/
@hmiron added a blog post for the Talos setup
https://www.beyondwatts.com/posts/setting-up-a-talos-kubernetes-cluster-with-talhelper/
After moving my homelab cluster from microk8s to Talos, I've written a new blog post outlining the cluster build:
Setting up a Talos kubernetes cluster with talhelper
https://www.beyondwatts.com/posts/setting-up-a-talos-kubernetes-cluster-with-talhelper/
@mmeier this is the English Channel which is ever so slightly warmer.. but not a lot!
The countryside is amazing in the Jurassic Coast. We are very lucky 👍
@mmeier you have the MONs also though?
@mmeier this is interesting - my Pi4s are running about 15% CPU just with the control plane. Temp is 50C but they are in the middle of the rack where the airflow isn’t great
Hey! Passing by to share my newest artwork, using the theme of something that I'm really enjoy to do, which is bikepacking.
Created @Blender and render done in Cycles.
The California Surf Map and other delights from Malin and Mizen #MapoftheWeek https://mapoftheweek.substack.com/p/mapping-the-surf
@hmiron really liking Talos. I haven’t written anything on setting it up yet but could do… I used talhelper 👍
@madduci time will tell… the UPS needs to go onto battery power to trigger the scaler (I could implement a timer before scaling gets triggered) and worst case, the cluster should scale down, get the ups online event and automatically scale back up!
Blog post now live
Creating a kubernetes autoscaling operator that responds to UPS events
This has been a really interesting project, thanks for all the feedback!
Mailbox.org 🇩🇪, a German e-mail provider, is running a special promotion:
New clients can enjoy six months of Premium or Standard plan for free. 😀
These plans include email, aliases, a calendar and contacts, online office, cloud storage, a custom domain, video conferences, XMPP web chat, team account & collaboration and digital legacy.
This offer is valid until May 31, 2025.
Check out the conditions here:
finally wrote something real about what i've been building!
too many of my infra workflows were buried in slack threads, docs, or shell history
so i started working on Atuin Desktop:
- runbooks that run
- local-first, crdt-powered
- embedded terminals, db queries, monitoring blocks
more words here: https://blog.atuin.sh/atuin-desktop-runbooks-that-run/
would love to know what you think! ❤️ :atuin:
@rachel I’ve always tried to shutdown down when the UPS kicks in. Here, I am only stopping workloads that need databases or PVCs. The rest (including the cluster and control plane) will ride the UPS down!
Of course, what’s really needed is to separate the cluster into two halves on separate UPS’s…. Perhaps with one half in the house and the other in the garage… But this will need to be another project!
Some final thoughts on this thread now I have recovered from the shock of it actually working:
NUT reports an UPS run time of 24 mins. It really means 6 minutes, even though the load does not increase dramatically when scaling down.
The low battery warning /shutdown message from the UPS occurs too late so now we scale down as soon as the power goes.
The time to scale down has been reduced from 3min30sec to about a minute by doing more in parallel.
@mmeier This got me thinking.... My HP EliteDesk Mini's have DisplayPort on the back and a USB-C connection on the front... Perhaps they could output the display over USB-C....
Alas no luck, they are probably too old for this but worth a try, it would have been sooo much easier!
Well that was an interesting one to debug... My blocky DNS service was down after a cluster restart
A given #metallb speaker won’t advertise the service if:
- the service has externalTrafficPolicy=local and there are no running endpoints on the speaker’s node
To use externalTrafficPolicy=local, the tolerations on metallb pods must match the tolerations on the destination pods
For once it wasn't DNS!