How LinkedIn extended ClickHouse from distributed tracing to metric discovery and analytics
In short
LinkedIn’s observability team has extended ClickHouse from distributed tracing to metric discovery and analytics, consolidating three systems into a single cluster that serves 150,000 queries per minute at 68 milliseconds average latency across 13 billion metrics. This migration off a legacy stack has reduced memory use to roughly one-fifth and compute to about two-thirds, with room to grow metric volume.
Key points
- LinkedIn reduced memory use to roughly one-fifth and compute to about two-thirds by migra…: LinkedIn reduced memory use to roughly one-fifth and compute to about two-thirds by migrating off a legacy stack.
- The observability team’s ClickHouse journey began with distributed tracing, which has sin…: The observability team’s ClickHouse journey began with distributed tracing, which has since expanded to include metric discovery and analytics.