How We Do Tracing at the Scale of Thousands of Services and Millions of Spans per Second
summary
A presentation about how Avito implements distributed tracing to handle millions of spans per second from several thousand services. It describes the problem tracing solves, the tracing architecture they built (including a service mesh that helps), and how that architecture processes the high volume of spans.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 5:27Tracing evolution at Avito
- 10:02Pipeline architecture
- 12:55Linking spans with Trace ID
- 16:50Sidecar proxy for tracing
- 19:00Sampling strategies
- 25:26High-cardinality data handling
- 27:56Tracing with logs
- 28:39Metrics with tracing
- 31:02Custom observability platform
- 32:48Debugging with traces
- 37:44Future plans
- 39:55Q&A
related talks
46mHow we at Avito analyze 5 million traces and conduct architectural oversight
49mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
22mDistributed Tracing
59mAn introduction to distributed tracing
57mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
37mObservability, Distributed Tracing & the Complex World
42mDistributed Tracing Opening the Door to Visualisation and Correlation
34mReduce Complexity with Distributing Tracing