
Host of @k8sfm, CNCF Ambassador, Freelance Content Creator. I help brands engage with and grow their audiences. https://xitter.cf/t.co/oKw0Shp7GS
Joined January 2013
- Tweets5.9K
- Following2.4K
- Followers4.7K
- Likes30.5K
"My biggest surprise was that kubelet pulls images serially by default. One big image blocks everything else."
Jan Ludvik on a one-flag fix for faster node startup
📺: ku.bz/B7TzKXyxf
I've had my ass kicked professionally, personally, and literally.
My boxing trainer weighs 118 kg. I'm 82. He's a former pro.
But between rounds, we dance.
Life guarantees we'll take hits. We might as well face them with grace, humor, and a little joy.
Can you trust an AI agent enough to let it touch production?
That’s the kind of question FLOCK ’26 is tackling in San Francisco on Oct 14 — real-world AI agents, SRE, platform engineering and autonomous production ops, without the usual hype.
Sign up- ku.bz/nh5ZFW_b9
Bart Farrell retweeted
Kubernetes, AI SRE agents, @birthmarkbart giving out cash prizes… and @aastaneh on the DJ decks 👀
Join us for LearnKube Day on Nov 9, Salt Lake City @ KubeCon
Free AI SRE crash course, talks, YAML Games + after-party.
Join in person or remotely: learnkube.com/learnkube-day-…
My boss can lift more than your boss- @danielepolencic in action!
Photo credit to @AmitJayShah, hope I can make it to the next meetup!
Kube Select starts now
An NGINX-to-Envoy migration produced five terabytes of metrics per second
The verdict: more telemetry, higher costs, and no better decisions
Henrik Rexed joins Salman Iqbal and Bart Farrell
ku.bz/848pmBN5_
With Dynatrace. Sponsored by LearnKube
"We need to keep the costs down. And more importantly, it's good for the environment. We need to keep our carbon footprint down."
Salman Iqbal on why scale to zero matters
📺: ku.bz/_hMPY1FGj
🗣️ Mike Stefaniak tackles how much Kubernetes knowledge developers should have. He advocates for a "middle ground" approach where platform teams build abstractions without completely hiding that applications run on Kubernetes
Watch: ku.bz/NH_jwkNcR
🎙️ What is Brandt Keller bringing to KCD New York?
We also have 10 free tickets — email [email protected] to claim one
🔗 ku.bz/JkjmffBzw
"CPUs are not real metrics."
Nicholas Eberts on why custom metrics beat CPU and memory for HPA scaling
📺: ku.bz/jlDL5XzCd
AI should assist the operators, not replace them
YongKang He on AI as a co-pilot for SRE
📺: ku.bz/8Q7Vy60P7
🗣️ Paul Butler explains why Custom Resource Definitions aren't always the best choice in Kubernetes
Watch: ku.bz/Dmn93dd7M
Bart Farrell retweeted
💥 The Kubernetes Automation Trust Gap study found that SREs and platform engineers trust optimization automation much less than engineering directors, even within the same organization. (Link to the report is in the comments.)
This makes sense when you consider who gets paged if something breaks. Directors see the savings dashboard, but SREs remember when a so-called safe recommendation reduced headroom just before a traffic spike, causing major issues.
One respondent said they would not trust the tool unless it can demonstrate that it understands the SLO impact and blast radius. Good savings numbers alone are not enough.
🗣️ Mike Stefaniak, Head of Product, Kubernetes and Registries at AWS, discusses why organizations keep AI tools in read-only mode for Kubernetes due to insufficient security controls and AI hallucination problems
Watch the full interview: ku.bz/PzjrglcZJ
🗣️ Andrew Charlton explains how instance matching addresses the limitations of StatefulSets when managing PostgreSQL clusters
Watch: ku.bz/fhZ_pNXM3
🗣️ How do you build SaaS workflows with Kubernetes Operators without turning one CRD into a dumping ground?
Alexander Held explains what went wrong with a 2,000-line CRD
ku.bz/TGy4Qn7Qs
🌟 LearnKube
🎙 🎙Bart



