Blog
Notes from our work. Published when we have something to say, not on a schedule.
Why we keep every backup in a second region
Backups that live next to the primary are not backups. Here is how our cross-region replication works and what it costs you.
Kubernetes 1.33 upgrade notes
All managed clusters were moved to 1.33 over the last two weeks. What changed, what we skipped and how to check your workloads.
A quieter on-call: alert routing done right
Fewer pages, better sleep. The three rules we use to decide what wakes an engineer up.
Egress pricing explained
Nobody likes surprise bandwidth bills. This is exactly how egress is metered on our platform.