delta-io / kafka-delta-ingest

A highly efficient daemon for streaming data from Kafka into Delta Lake
Apache License 2.0
359 stars 79 forks source link

Add lag reporting #77

Closed xianwill closed 3 years ago

xianwill commented 3 years ago

This PR adds lag reporting to kafka-delta-ingest via statsd metric emissions.

Two flavors are added - Buffer lag and write lag. Buffer lag represents the last offsets stored in buffer. Write lag represents the last offsets written to the target delta lake table. We report four metrics for each:

Buffer lag reports roughly every minute (but requires a message receive to trigger). Write lag reports every time a write is performed.