LodeHQSubscribe →

Slack multi-cloud AI, AWS Resilience Hub, Mercari Cloudflare

DevOps · 2026-05-30

Containers & Orchestration
Azure Container Registry caches only requested architecture manifest, then fires webhooks on async copy completion11 MIN

When a multi‑architecture image is pulled through ACR Artifact Cache, only the manifest list and the specific architecture’s manifest are cached, while other architectures remain upstream until requested. The cache then asynchronously copies these artifacts, and ACR push webhooks fire once the copy finishes, indicating the image is fully stored locally and storage charges start.

Observability & Reliability
AWS Resilience Hub adds AI‑driven failure analysis and org‑wide policy enforcement5 MIN

AWS launches the next‑gen Resilience Hub, letting SREs define modular resilience policies, map business‑critical paths, and run generative AI‑powered failure‑mode assessments. Integrated with AWS Organizations, it auto‑discovers dependencies and provides enterprise‑wide reporting to validate compliance across hundreds of accounts.

Why Reliability Metrics Age Faster Than the Systems They Measure10 MIN

A senior SRE details how SLIs and reliability dashboards become misleading as services evolve, with green screens masking real customer issues. The post outlines recurring failure patterns across major observability stacks and stresses the need for continual metric reassessment to keep reliability signals accurate.

Cloud & Platform Engineering
Mercari Migrates Legacy Image Pipeline to Cloudflare Images with Zero‑Downtime Dual‑Path Rollout9 MIN

Mercari's Platform Network team moved a long‑standing image resizing service to Cloudflare Images while keeping all existing URLs alive by running legacy and new paths in parallel and using Cloudflare edge origin overrides with S3 host‑header HTTPS. The post details traffic estimation, safe rollout tactics, and cost trade‑offs for a seamless migration.

Slack launches multi-cloud AI platform spanning AWS and GCP16 MIN

Slack re‑engineered its AI serving stack over three years, moving from AWS SageMaker to Bedrock, then to a hybrid architecture across AWS and Google Cloud. The new multi‑cloud setup aims to cut scaling latency, avoid GPU shortages, and keep pace with the latest LLM advances, targeting full deployment by early 2026.

DevSecOps
ISO‑27001 Compliance as Code: Terraform‑Driven Security on AWS11 MIN

A Terraform‑first AWS startup built ISO‑27001 certification by codifying infrastructure, access controls, encryption, monitoring, and vulnerability management. This allowed audit evidence to be generated directly from Terraform definitions, ensuring continuous compliance and faster enterprise sales. The post details the process and lessons learned.

Get DevOps in your inbox, every issue.
Subscribe free
Privacy · Terms · About · Contact
© 2026 LodeHQ