← All work
Pulse
Event pipeline and analytics API processing 2B events a month with sub-second queries.
The challenge
An analytics startup’s pipeline buckled past a few hundred million events a month and dashboards took minutes to load.
What we built
- 1
Re-architected ingestion with partitioned streams and columnar storage.
- 2
A query API with caching that keeps common dashboards under a second.
- 3
Anomaly alerts on pipeline health with clear runbooks.
- 2B
- events / month
- <1s
- p95 queries
- −41%
- infra cost
Stack
- Go
- ClickHouse
- Kafka
- Kubernetes
- AWS
Sample case study. Swap in real screenshots, numbers and a client quote before launch.