DevOps / SRE
11 entries in DevOps / SRE.
Showing 11 of 11 · sorted by a–z
Backup Restore Drill
The rehearsed, timed exercise of actually recovering data from backups — because an untested backup is a hypothesis, not a safety net.
Chaos Engineering Practice
The deliberate injection of controlled failure into production systems to discover the weaknesses that only surface under stress — turning fear of the unknown into an engineering discipline.
Engineering Capacity Planning
Forecasting demand against infrastructure headroom — in business units, not just CPU — so the platform survives its busiest hour without paying for the busiest hour all year.
Ephemeral Preview Environment
A short-lived, per-branch or per-pull-request deployment that lets reviewers see and test changes in isolation — the practice that quietly cuts review cycles in half.
Error Budget Policy
The explicit, negotiated agreement between engineering and product that says what happens when reliability drops — the mechanism that turns SLOs from posters into decisions.
Golden Signals Monitoring
The four service-level metrics — latency, traffic, errors and saturation — that together tell you almost everything you need to know about a running system.
Incident Commander Role
The single named coordinator who runs a major incident — directing responders, owning communication and making decisions — so the best engineers can fix the problem instead of chairing a forty-person call.
Load Shedding and Backpressure
Deliberately refusing or slowing work so a system degrades gracefully instead of collapsing — the difference between serving 90% of users and serving none.
Progressive Delivery
The practice of releasing changes to production in controlled, observable stages — a small percentage of users first, then wider audiences as confidence grows — rather than to everyone at once.
Schema Change Discipline
The expand-and-contract method for evolving a live database — add the new shape, migrate in steps, remove the old shape later — so deploys and rollbacks never break a running system.
Secrets Rotation Policy
The rules and machinery for changing credentials on a schedule and on demand — so a leaked key has a short useful life and rotation is not an outage.