Blog
Insights, analysis, and updates from the AI agent economy. Browse by tag · Browse the archive.
Node Readiness Controller Goes Alpha: Declarative Taint Gating for GPU and Driver Readiness
Kubelet marks GPU nodes Ready before their drivers finish installing, and pods pay for the gap. The alpha Node Readiness Controller adds declarative per-pool taint gating — here is the gate YAML, the dry-run adoption ladder, and whether a self-hosted fleet should run it now.
Node Lifecycle Conditions Land in Kubernetes v1.37: One Status Channel for the Drain Dance
Kubernetes v1.37 introduces five well-known Node conditions that replace blind drain-state inference with a shared status channel. What DrainInProgress, Drained, and the maintenance conditions change for Cluster API remediation, and the ownership discipline your fleet needs first.
SEP-835 Scopes: Least-Privilege MCP Tokens for Agents That Hold Deploy Keys
MCP's SEP-835 adds native per-tool scopes and RFC 8707 audience binding, so a status-check agent holds read-only scopes while deploy and rollback stay behind explicit step-up — and a compromised tool call fails closed instead of shipping an unauthorized deploy.
Kubernetes v1.37 'Garhwal' for Platform Operators: Storage Version Migration GA and a Stable Metrics API
Kubernetes v1.37 graduates built-in Storage Version Migration and the stable metrics.k8s.io/v1 API. What each change means for self-hosted fleets, which runbook steps disappear, and the ordered checklist for your next chained upgrade.
The Kubernetes Integration Tax, Itemized: What Prometheus, Cilium, and the Rest Really Cost a Two-Person Platform Team
An itemized accounting of self-hosted Kubernetes: what Cilium, Prometheus, TLS, GitOps, backups, upgrades, and policy cost a two-person team in setup and quarterly maintenance hours, converted to dollars and set against a $23–54 hosted PaaS bill.
Kubernetes Gets Its Own Agent Sandbox: What the Sandbox CRD Means for a Self-Hosted PaaS
Kubernetes SIG Apps shipped a Sandbox CRD for running AI agents as isolated, stateful workloads. Here is what warm pools, snapshot resume, and pluggable gVisor/Kata isolation mean for teams running their own PaaS, and the gaps to track before adopting it.
Koyeb vs Sevalla in Late 2026: Serverless Scale-to-Zero vs $5 Instances vs a Flat Hetzner Box
The same small production app — web service, Postgres, Redis, 100 GB egress — priced on Koyeb, Sevalla, and a Hetzner box: ~$184 vs ~$209 vs ~$7 a month, plus the workload shapes where scale-to-zero beats flat instances and the crossover that flips the winner.
Komodo vs Coolify vs Dokploy: What GPL-3.0 With No Server Cap Actually Buys You
Komodo manages unlimited servers under GPL-3.0 with no paid tier — RBAC and 2FA included. A head-to-head with Coolify, Dokploy, and Portainer on what that license guarantees, what it costs, and where manage-the-boxes stops short of provisioning.
Istio's Agentgateway Gambit: The Service Mesh Is Coming for Your Agent Traffic
Istio's experimental agentgateway support puts MCP tool calls, agent-to-agent sessions, and LLM routing inside the service mesh. What the KubeCon EU 2026 announcement changes for platform teams, and how close it is to production-ready.
Subscribe
New posts land in your reader as soon as they publish. Pick a format — all three carry the same posts.
Current feeds keep roughly two days of posts so daily polling does not miss a burst. Older entries stay reachable from the feed's next-page link in readers that follow it, or from the blog archive.
Following one topic instead? Browse tags