KIP-613 — Add end-to-end latency metrics to Streams
Accepted Kafka 2.6 StreamsMetrics
Introduces an `end-to-end-latency-avg` / `end-to-end-latency-max` task-level metric in Kafka Streams that measures the time from when a source record's timestamp was created (its event time) to when it is processed. There is no existing metric that captures the full pipeline latency from event time to processing time, making it difficult to bound or tune real-time application SLOs.
Details
| Author | A. Sophie Blee-Goldman |
| Status | Accepted |
| Kafka Version | 2.6 |
| JIRA | KAFKA-9983 |
| Wiki | View on Apache Wiki |
| Created | 2020-05-11 |
| Last Modified | 2020-07-15 |
Explore how this KIP affects the Kafka protocol in the Protocol Explorer, or see the full KIP database.