TLDR DevOps 2026-03-18

Legacy search tools were never designed for the way enterprise data actually works. OpenSearch delivers AI-powered vector retrieval, agentic workflows, and synthesized answers from the unstructured data already in your pipeline. Open source, Apache 2.0 licensed, and backed by 3,000+ contributors across 400+ companies. No lock-in, no licensing surprises.

[See how OpenSearch handles what your current search can't]

Microsoft launched a preview of its Remote Azure DevOps MCP Server, a hosted version that eliminates setup requirements and works with Visual Studio and Visual Studio Code through streamable HTTP transport. The remote server requires Microsoft Entra authentication and doesn't yet support other tools like GitHub Copilot CLI or Claude Desktop, though Microsoft plans to eventually phase out the local MCP Server in favor of this cloud-based option.

DigitalOcean's App Platform Skills is a collection of open-source AI-native playbooks that give coding assistants like Claude, Copilot, and Cursor up-to-date knowledge about deploying applications to production environments, including 12 specialized skills covering everything from greenfield app design to Heroku migration. The system uses opinionated defaults based on thousands of real deployments—like VPC networking by default and PostgreSQL v16 with SSL required—and keeps credentials secure by having AI assistants reference GitHub Secrets by name rather than handling sensitive values directly.

Monitoring Karpenter Prometheus metrics reveals autoscaling behavior, provisioning latency, consolidation activity, and cloud provider issues. Tracking scheduling, disruption, controller, and cost metrics helps diagnose scaling delays, API throttling, and inefficiencies affecting Kubernetes performance and cost.

Each additional layer of review in a process introduces massive latency—often slowing delivery by an order of magnitude—because most time is spent waiting rather than doing, and tools like AI don't meaningfully fix this bottleneck. The real path to speed and quality is not adding more reviews, but eliminating them through better system design, small trusted teams, and continuous root-cause improvement so issues are prevented rather than caught late.

An Octopus project deploys Kubernetes YAML to a mock Kubernetes API using the octopussolutionsengineering/k8s-mockserver execution container, enabling testing without a real cluster. The setup includes a Kubernetes target, SBOM vulnerability scanning, environment-based step conditions, and a daily security scan trigger.

Datadog engineers reduced the Go based Datadog Agent's binary size after it grew from 428 MiB to 1.22 GiB by auditing dependencies, isolating optional code with build tags, minimizing reflection, and avoiding plugins, significantly cutting bloat without removing features.

AWS introduced a GitHub Actions “Deploy Express Service” action that automates CI/CD for Amazon ECS Express Mode by building Docker images, pushing them to Amazon ECR, and deploying updates on code commits. The pipeline uses OIDC authentication with IAM roles for secure, temporary AWS access.

Python 3.15's JIT has finally achieved modest but real speedups (≈5–12% on average) after earlier versions underperformed, thanks largely to community-driven development, better contributor onboarding, and a few high-impact technical bets like trace recording and reference count optimizations. The turnaround was less about any single breakthrough and more about the right people, collaboration, iterative improvements, and some luck aligning at the right time.

AI agents can generate code quickly, but improving it requires real production feedback. Join Honeycomb on April 1 to see how engineering teams use observability data to help AI agents produce better, more reliable code. Secure Your Spot

The Kubernetes team rewrote kpromo, a critical image promoter tool that handles every container image published to registry.k8s.io, reducing the codebase by 20% while cutting production promotion jobs from 30 minutes to just 2-15 minutes across different phases.

A new production guide for Kubernetes 1.35 clarifies the critical difference between pod restarts and recreations: if the pod UID changes, it's a complete recreation with restart counts resetting to zero, but if the UID stays the same, it's just a container restart within the existing pod.