0.5.5

Bug Fixes:

• [37] Support non-string values in consumer.properties files (thanks @jaewonkim)

0.5.4

Bug Fixes:

• [30] Exclude kafka-python versions 1.4.5 and 1.4.6 due to performance issues.

0.5.3

Enhancements:

• [35] Change __consumer_offsets key parsing error logs to warnings (thanks @jutley)

Bug Fixes:

• [33] Fix exporter offset to track the offset of the next message, not the current one, to make offset/lead/lag metric semantics consistent (thanks @jutley)

0.5.2

Bug Fixes:

• [29] Remove metrics for deleted consumer groups (thanks @jutley)

0.5.1

Bug Fixes:

• [28] __consumer_offsets messages from Kafka 2.1+ not supported

0.5.0

Breaking Changes:

• kafka_consumer_group_commits metric renamed to kafka_consumer_group_commits_total, due to change in Prometheus client

Enhancements:

• [27] Add kafka_consumer_group_commit_timestamp metric to track time of the last commit from a consumer group (thanks @jutley)

Bug Fixes:

• [15] --bootstrap-brokers default value shouldn't override value in consumer config file (thanks @jsargiot)
• [18] Fix memory leak with Python 3.7
• [21] Don't let lag metrics go below 0
• [22] Handle exception parsing strange __consumer_offset messages (thanks @jutley)

0.4.0

Enhancements:

• [12] Rework to not require .schedule() method on clients from kafka-python 1.3.x, adding support for 1.4+

0.3.1

Bug Fixes:

• [12] Require kafka-python 1.3.x, 1.4+ drops the .schedule() method on clients

0.3.0

Enhancements:

• [9] Add low-water mark metric (thanks @rdonnelly3)
• [8] Add lag and lead metrics

0.2.0

Breaking Changes:

• Default port changed from 8080 to 9208, see here for why

Enhancements:

• Add high-water mark metric (allows calculating lag)
• Support configuring the consumer via properties files
• Allow setting arbitrary logging levels