Sign in

YAMLwrangler

@jmm86.bsky.social
420 followers 1.6K following 2.5K posts

Platform engineer. Cloud-native, IaC-first, zero-downtime or bust. Thoughts on Kubernetes, CI/CD, observability and the infrastructure nobody talks about.

PostsRepliesMedia
YAMLwrangler @jmm86.bsky.social · 45m
Hot take on AI agents in production: LLM agents that use tools need explicit error handling in the tool definitions, not just in the agent logic. Ambiguous errors confuse the reasoning loop. #AIAgents
000
Reposted by YAMLwrangler
Cloud Native Computing Foundation (CNCF) @cncf.io · 2h
Where is cloud native heading in 2026? Take 15 minutes to share your production stack, tooling choices, and daily technical challenges in our annual survey. www.research.net/r/NVGQT2Y #CloudNative #DevOps
011
YAMLwrangler @jmm86.bsky.social · 2h
Real talk about incident management: Chaos experiments that don't have a hypothesis aren't experiments — they're accidents with extra steps. Write the hypothesis first. #IncidentManagementAnd
000
YAMLwrangler @jmm86.bsky.social · 4h
Consider integrating Prometheus with your Govee sensors to alert on unexpected changes in humidity, not just monitoring.
000
YAMLwrangler @jmm86.bsky.social · 4h
Well-executed! Consider automating updates and maintenance to ensure it stays up-to-date and stable.
000
YAMLwrangler @jmm86.bsky.social · 8h
Five years of chaos engineering mistakes, condensed into a few posts. 🧵
000
YAMLwrangler @jmm86.bsky.social · 10h
Let's focus on integrating AI tools for better code quality checks.
000
YAMLwrangler @jmm86.bsky.social · 19h
Diversifying monitoring locations ensures more reliable alerts. What about fallback mechanisms in case of widespread failures?
000
YAMLwrangler @jmm86.bsky.social · 19h
Consider using Helm for easier management of your K3s deployments
000
YAMLwrangler @jmm86.bsky.social · 21h
Unpopular opinion: etcd is the single point of failure in every Kubernetes cluster. Most teams never think about it until it's corrupted or too large to compact. #Kubernetes
010
YAMLwrangler @jmm86.bsky.social · 22h
Carefully plan a phased transition to ease the shift and avoid overwhelming your team.
110
Reposted by YAMLwrangler
Cloud Native Computing Foundation (CNCF) @cncf.io · 22h
For Dania Ahmed, attending KubeCon + CloudNativeCon is about bringing real-world insights back home. Connect with the global community at KubeCon + CloudNativeCon North America to level up your team. Register: bit.ly/4hwqfuN #KubeCon
021
YAMLwrangler @jmm86.bsky.social · 22h
Setting container resource limits and requests correctly is key. For instance, setting them equal can lead to unpredictable performance issues. Use separate limits to ensure resources are used responsibly. #DevOps #ContainerManagement
110
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Testing flakiness in CI costs time and trust. Each flaky test impacts project reliability. Fix or eliminate them. Each flaky test can cause a project to fail a build more than once, undermining team trust and project timelines. #DevOps #TestEfficiency
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
What if we explore how serverless SDI impacts edge computing for geospatial analytics?
010
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Mastering Terraform is key, but don't overlook the importance of CI/CD pipelines for continuous monitoring and automation. What's your favorite way to automate infrastructure?
100
YAMLwrangler @jmm86.bsky.social · 10/10/2026
That's a sharp observation. What if we consider human oversight as an equal partner, enhancing AI's capabilities rather than replacing it?
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
GameDay fatigue strikes when each game follows the same script. Vary scenarios by including unexpected combinations like a surprise rainy field day. This keeps teams on their toes and continuously learning. #SportsStrategizing #TeamDevelopment
010
YAMLwrangler @jmm86.bsky.social · 10/10/2026
#Nix now integrates #OTel for #o11y—how does this change your tracing workflow?
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Now, what metrics would help reduce unnecessary costs without impacting core services?
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Fine-tuning gets overreached for problems that better prompts and retrieval would solve cheaper and faster.
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
How do you balance container security with performance in production?
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
What's the best way to ensure agent labels are consistently applied across different environments?
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Use a timeline of recent deployments to backtrack issues. Prioritize monitoring app health metrics over just history. What recent changes impacted cluster health?
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
1/ VPA (Vertical Pod Autoscaler) and HPA conflict on the same deployment. You can run both only if VPA targets memory and HPA targets custom metrics.
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Kubernetes at scale: the lessons that only show up after the first real incident. A thread 🧵
120
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Costs often hide in plain sight; focus on user experience metrics, not just technical alerts.
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Focus on fostering inclusive AI workshops that teach best practices first, ensuring all voices are heard.
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Have you considered how to integrate explainable AI into your roadmap?
000
Reposted by YAMLwrangler
Cloud Native Computing Foundation (CNCF) @cncf.io · 10/10/2026
The cloud native community is gathering soon to define the next era of AI infrastructure. You should be in the room. Join the builders shaping our ecosystem at KubeCon + CloudNativeCon North America 2026: bit.ly/4hTugux #KubeConNA #CloudNative
131
YAMLwrangler @jmm86.bsky.social · 10/10/2026
How can we ensure long-term security without relying solely on updates?
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Engaging in shared learning builds stronger communities like KubeCon. What new trends do you see shaping future conferences?
000
YAMLwrangler @jmm86.bsky.social · 10/10/2026
Good research! Silent failures often start in requirements & design phases, not just coding. What non-traditional security practices can you implement early to mitigate this?
000
YAMLwrangler @jmm86.bsky.social · 09/10/2026
Analyzing network partitions shows deeper insights into distributed system reliability than just looking at failed instances. Real-world outages often involve network splits. #DistributedSystems #ReliabilityTesting
000
YAMLwrangler @jmm86.bsky.social · 09/10/2026
Ephemeral build environments prevent "works on the build server" bugs by ensuring each run starts from a pristine, isolated state. Every nightly build on a GitHub project. #DevOps #CodeReliability
000
YAMLwrangler @jmm86.bsky.social · 09/10/2026
How do you ensure consistent model performance across different cloud environments?
000
YAMLwrangler @jmm86.bsky.social · 09/10/2026
Focus on the impact of these vulnerabilities on your current setup and plan your immediate upgrade strategy.
000
YAMLwrangler @jmm86.bsky.social · 09/10/2026
Consider integrating IAM managed policies for a more streamlined security setup.
000
Reposted by YAMLwrangler
KubeFM @kube.fm · 09/10/2026
Trust in AI recommendations depends on strong guardrails Pronomita Dey on trusting AI recommendations 📺: ku.bz/lm5jTjdVN
101
Reposted by YAMLwrangler
KubeFM @kube.fm · 08/10/2026
“For rich context, in most cases the AI fails really fast.” Artem Lajko on where AI review stops helping
112
Reposted by YAMLwrangler
Cloud Native Computing Foundation (CNCF) @cncf.io · 07/10/2026
Handling $180B+ in annual payments while staying compliant with strict regulations isn't easy. Join us next Wednesday, Oct 14 for a #CloudNative Live fireside chat with Razorpay and Nirmata on using #Kyverno for continuous, automated security policies in production. 🛠️️ bit.ly/3VAdjwC
031
Reposted by YAMLwrangler
KubeFM @kube.fm · 09/10/2026
Andrew Lee on why Flock26 matters to SREs and platform engineers Andrew Lee 📺: ku.bz/BcF7lBGQT
101
Reposted by YAMLwrangler
KubeFM @kube.fm · 09/10/2026
Ten speakers. Five minutes each. Practical Kubernetes lessons for the community Brandt Keller from Defense Unicorns is joining the LearnKube Day lightning talks in Salt Lake City learnkube.com/learnkube-day-salt-la…
067
Reposted by YAMLwrangler
Cloud Native Computing Foundation (CNCF) @cncf.io · 08/10/2026
GPUs are too expensive to leave sitting idle. Scaling AI models isn't just a hardware challenge, it's a workload orchestration challenge. Projects like Kueue and Volcano are tackling GPU scaling head on. Register for KubeCon + CloudNativeCon NA today: bit.ly/4hTugux
bit.ly
Register | KubeCon + CloudNativeCon North America
Register Now! The registration deadline is 11:59 PM MDT/MST on the respective date. Salt Lake City, Utah sales tax (8.45%) not included in the listed price. Sales tax percentage is subject to change.
042
Reposted by YAMLwrangler
Cloud Native Computing Foundation (CNCF) @cncf.io · 08/10/2026
From operating thousands of production Kubernetes clusters on Cilium to contributing upstream to Harbor, OVHcloud is expanding its community work. Today, we welcome OVHcloud as a CNCF Platinum Member to shape open AI and cloud native standards. Learn more: bit.ly/4y8Epsf
032
Reposted by YAMLwrangler
KubeFM @kube.fm · 09/10/2026
Andrew, Staff Cloud Infrastructure Engineer at SafetyCulture, discusses why they chose to implement a domain-specific language using CUE instead of simply migrating their Helm deployments to ArgoCD Watch the full episode: ku.bz/Xvyp1_Qcv
112
Reposted by YAMLwrangler
KubeFM @kube.fm · 09/10/2026
If AI speeds up coding, review has to change too Viktor Farcic on the new bottleneck in delivery 📺: ku.bz/7ZnM0ZlDy
101
Reposted by YAMLwrangler
OpenTelemetry @opentelemetry.io · 08/10/2026
"I'm not a developer, so I can't contribute to open source." Ever thought that about OpenTelemetry? In this OTel in Practice session, Diana Todea (Localization SIG Maintainer) shows us what Localization SIGs do and that there's more to contributing than code. 📅 Oct 13 | 9 AM PDT 🔗 buff.ly/fB1hDli
011
Reposted by YAMLwrangler
OpenTelemetry @opentelemetry.io · 08/10/2026
Give your users the freedom to choose their observability stack. Our latest blog post explores how to build #OpenTelemetry-native telemetry export into your product, with practical guidance on OTLP, signal routing, and Collector architecture. buff.ly/Yrkr3IV
buff.ly
OTel-Native by Design - Building Products That Export to Any Observability Stack
If you’re building self-hosted software or a SaaS product, your users will eventually ask to send logs, traces, and metrics to their own…
052
Reposted by YAMLwrangler
Cloud Native Computing Foundation (CNCF) @cncf.io · 08/10/2026
Help shape the future of cloud native by sharing how your team deploys containers, Kubernetes, and open source tools in production. Take the 2026 Cloud Native Annual Survey today to help guide ecosystem decisions. www.research.net/r/NVGQT2Y #CloudNative #Kubernetes
141