Replying to
@kauer@aus.social @oberonsghost@aus.social It was reported on the BBC so decent chance it did actually happen: https://www.bbc.co.uk/news/newsbeat-36580448
Remote
LiquorVicar
@LiquorVicar@mastodon.social
mastodon 4.8.0-nightly.2026-10-06I don't know what I'm talking about.
If you find me where I live in the UK (NW England), I may well be ranting about software development, sports-ball (ITFC, cricket, cycling, "Go Bears!" etc). I'm probably harmless but you might want to walk away slowly backwards just in case.
0 Followers
0 Following
17 Posts
Joined November 11, 2017
Website:
Replying to
@imalcolm@mastodon.social @swisslet@mas.to It's vital to know there are no sun spots on your frozen dairy confection.
Replying to
@europlus@social.europlus.zone @davidgerard@circumstances.run Is the Tolerance paradox a potential parallel here too?
Replying to
@GeePawHill@mastodon.social it's also used as the soliloquy at the end of Withnail and I IIRC.
Replying to
@DJDarren@mendeddrum.org @GeePawHill@mastodon.social Oh yeah, thinking about it I haven't watched it in an age...
Fun fact, after we moved into our current house we found out that it had been named "Marwood" at some point in the past.
Replying to
@jess@blop.social 108. 108! 108!! I'm a big cheese fan but I've never curated a board that big. I bow down to your cheesy awesome-ness (and next time you do it, can I get an invite 😉 )
Replying to
@kauer@aus.social @oberonsghost@aus.social Came here to post the same story. No idea if it's true I'll have to check. But it's such a great story.
Any #k8s #kubernetes experts able to give me a bit of a second opinion? We have a DOKS (Digital Ocean) cluster using Cilium/Envoy/GatewayAPI. We're hitting an issue where the Envoy config is getting to a certain size (~4Mb in total, around 500 non-control plane pods) and we're getting overflow errors. At this points pods and associated resources are all created fine but none of the newly created pods are accessible externally (older pods are fine). Is this a known limit?
Replying to
@Madic@chaos.social We can be running ~510 (I can't remember the exact number offhand) app pods and it's fine. We add one more and these warnings start appearing (and the extra pod is not routable externally). If I then delete that pod, the warnings stop. At no point have I seen the existing pods affected.
Replying to
@Madic@chaos.social Looks like it's a constraint of using Envoy Proxy at this scale. Digital Ocean have suggested we switch to Traefik for ingress and we're getting much better results (I had 1500 pods running earlier with no problems).
Insect
Replying to
@joannaholman@aus.social I used to cycle past it regularily when I lived in Oxford! Legend has it that Dragon Hill, just below the horse, is where St George slew the dragon.
Replying to
@Madic@chaos.social I haven't seen those messages. The warnings we're getting (like several of these every second) are:
"[StreamEndpoints gRPC config stream to xds-grpc-cilium closed since <*> ago: 14, upstream connect error or disconnect/reset before headers. reset reason: overflow" module=agent.controlplane.envoy-proxy subsys=envoy-config
Replying to
@Madic@chaos.social I haven't migrated our prod cluster yet but tests using traefik in my dev cluster are very encouraging. Scaled to 1500 pods easily and might go up to 2500 or so (before I hit the limit on NodePorts). We've set a 5s throttle duration as we add pods in batches of up to 400.
Replying to
@ericwastl@hachyderm.io Thank you so much for all the effort you've put in over the years. Fully respect your decision to make these changes. Your sanity is important!