📰 Dev.to · DevOps Oasis

Docker version tags help people, but production deployments should use image digests. Our team’s rollback evidence, limits, and policy shift.

A Tuesday IAM cleanup exposed a stale Terraform state lock, an overlooked S3 permission, and why we refuse opportunistic provider upgrades.

Five scrum myths tested against a platform team’s spillover, status meetings, story points, and one manager saying no to Sales.

Build and test Kubernetes NetworkPolicy rules that allow frontend traffic, PostgreSQL access, and DNS while blocking everything else.

Our team answers awkward DevOps questions after a failed release handoff, covering ownership, production access, pager duty, metrics, and rollouts.

One year after replacing sprints with kanban, our nine-person platform team grades the move B: clear limits helped, but board upkeep became real work.

Six leadership anti-patterns from our platform team: approval bottlenecks, hero culture, bad paging routes, misleading metrics, and unpaid on-call.

A practical beginner primer on SRE, using our payment API alerts, SLOs, error budgets, and on-call decisions from one difficult day.

We tested five common Ansible claims during a 28-host Debian 12 cutover, including check mode, roles, idempotence, and deployment speed.

Jenkins, GitHub Actions, and GitLab CI compared after a 43-minute build queue, with concrete picks for small, growing, and platform teams.