Jens Rantil

See https://prometheus.io/docs/practices/naming/#base-units. https://github.com/Southclaws/restic-robot/blob/0ba420b9d671b306d273a4e28bf9bc19bc34af9d/metrics.go#L30 doesn't adhere to the base unit best-practice.

Contents: 1. Please give an example of the Kafka output. 2. Preferably, write a test for the output to parse it. Would help in cases like #51.

See https://github.com/apache/kafka/pull/4271.

This would help in debugging why scraping is taking so long in #41.

https://github.com/kawamuray/prometheus-kafka-consumer-group-exporter/blob/master/prometheus/collector.go#L74

Add process monitoring

1

Many exporters (such as [HAProxy exporter](https://github.com/prometheus/haproxy_exporter/blob/master/haproxy_exporter.go#L455)) supports a PID as CLI argument to also monitor the OS process. This is useful for example to see CPU usage and/or memory usage....

Document in README that only new consumer API is supported (at the moment - PRs welcome!). See #14.

Discuss with upstream if output can be standardized

2

...or alternative ways to avoid having to parse the output. Are there Go clients we can use to extract the same group information, perhaps?

Does incremental repair using ranges really work?

2

https://groups.google.com/d/msg/nosql-databases/peTArLfhXMU/IuUYnnUhBgAJ claims it doesn't work. I was investigating this because I was curious if I could slowly enable incremental repairs using this script.

Document in README how to install this

2

Document in README how to install this.