Profile for taisph

Header for taisph
Avatar for taisph
Display name
Tais Hansen
Username
@taisph@me.aedifex.dk
Role
admin

About taisph

Bio

😴 Dad
🤓 Senior DevOps Engineer
🤠 Docker/Kubernetes/GCP wielder
🐹 Gopher (the emoji is definitely a gopher if you squint a bit)
😎 Gamer

Stats

Joined
Posts
97
Followed by
13
Following
49

Recent posts

I use #YouTube Music to listen to music *gasp* and not recorded radio shows with promo talk-overs.

YouTube Music adds these "albums" to the auto playlists and there's no apparent way to stop it other than disliking the "album" every time, which is kind of sad since I usually like the artist(s) - just not the talk-over parts. Same reason why I don't listen to regular radio. Too much talking.

One solution could be to reclassify these types of "albums" as podcasts and keep them out of the auto playlists.

Open thread at this post

Just helped my kid set up an Elgato Stream Deck on his Ubuntu system. Went a lot more smooth than I had feared with only a little bit of guesswork around user/group access (user must be in the plugdev group on Ubuntu). UI is intuitive and let him easily configure all the buttons he had planned.

Thank you to the people involved in the project!

https://github.com/streamdeck-linux-gui/streamdeck-linux-gui

Open thread at this post

Just upgraded my #opentelemetry collectors to v0.97 and now the "otlp_exporter_seen_total" metrics from all the javaagent clients are being split up over multiple metrics with between 0 and 17 "_total" strings appended. Good times.

Now to figure out if this is caused by the javaagent or the collector.

Open thread at this post

Did the #Ubuntu #Gnome terminal get a rollback of sorts in Mantic? It was working perfectly with window tiling sizing the window to full half of the screen. Now it once again locks window size to fit the content as with Jammy, leaving ugly oddly sized gaps around the window. 😭

Open thread at this post

Operating Kubernetes is kind of like a game of Tetris. Pieces (pods) of varying sizes (resource requests) needs to fit snugly into the allotted spaces (nodes). If a piece is placed (scheduled) badly you won't be able to utilize the space as well and in a worst case scenario can't add the last few pieces (unscheduled pods).

Open thread at this post

I cannot stress enough how important it is to exclude autoscaler resource updates in ArgoCD.

After the GKE 1.25 upgrade, the application controller CPU usage went through the roof. Gettting ArgoCD upgraded to get the ignoreResourceUpdates feature and excluding autoscaler resource updates got the CPU usage to drop below even pre-upgrade usage.

Open thread at this post