Blog
Notes from the Quillmetric team
How we build and operate a large-scale observability platform — and what breaks along the way.
Engineering
Tail-based sampling in production
Keep the traces that matter and drop the boring ones — how we cut trace storage by 90% without losing a single errored request.
March 12, 2024 · 8 min read
Engineering
Taming metric cardinality
A label per user_id feels harmless until it isn't. A practical guide to high-cardinality metrics that won't blow up your bill.
February 27, 2024 · 6 min read
Incident
Postmortem: the eu-central rebalance
A 34-minute latency spike, one very hot shard, and the dashboard that told us exactly where to look.
January 18, 2024 · 5 min read