Kubernetes backup: what to actually look for
"Which backup tool is best?" is the wrong first question. Resource backup, CSI snapshots, filesystem backup and a tested restore path — chosen in that order.
Read the post →Blog
Engineering notes from the platform.
Practical, engineer-written notes on Kubernetes, observability, platform automation and AI operations — no fluff, no vendor pitches.
All posts
Scaling Prometheus metrics with Grafana Mimir
When metrics storage grows from gigabytes to terabytes, it stops being a monitoring problem and becomes an architecture problem — without throwing away Prometheus.
Read the post →Where eBPF fits in Kubernetes observability
Node Exporter, OpenTelemetry and eBPF aren't competing tools — they observe different layers. What eBPF actually adds, and where it doesn't replace instrumentation.
Read the post →Zero-downtime Kubernetes upgrades: a practical playbook
What actually breaks during a cluster upgrade, and the rollout sequence — PodDisruptionBudgets, surge order, and validation gates — that keeps traffic flowing through it.
Read the post →Taming observability cost at scale
High-cardinality metrics and verbose logs don't have to mean runaway bills. Cardinality control, tiered retention and smarter routing, explained.
Read the post →Where AI actually helps in incident response
A grounded look at what LLM-assisted investigation is good at today — timeline assembly, change correlation, log triage — and where humans still decide.
Read the post →Let's engineer your platform.
Tell us what you are running today.Tell us where you want to go.
We'll help design the path between them.