Also published on the CNCF blog. Cross-post with canonical link to the CNCF version. The 3...
Also published on the CNCF blog. Cross-post with canonical link to the CNCF version.
The 3 AM Call
We got paged one Tuesday morning. A critical production service had crashed under traffic—not gradually degraded, but crashed. Hundreds of pending pods. Users were seeing 15–20% error rates. The incident postmortem was brutal: reactive autoscaling had fired, but it was already too late.
The timeline looked like this:
06:00 Traffic spike arrives
How we added native Spanner support to Kubernetes autoscaling — and what we learned along the...
If you run vLLM, Triton, or any other inference server on Kubernetes, you have probably noticed that...
💥 The 3:00 AM On-Call Nightmare If you’ve ever carried an on-call pager for a production Kubernetes...
Introduction AI workloads are exploding. Enterprises are now scaling distributed training...
Co-authored with Netanel Kadosh You've built a great inference service. Auto-scaling is configured....
GPU scheduling, MLOps platforms, and platform engineering strategies for AI workloads on Kubernetes in 2026. Optimize utilization…
