Redpanda Minion is a feature-rich and flexible Prometheus Exporter to monitor your Redpanda or Kafka cluster. All valuable information that are accessible via the Kafka protocol are supposed to be accessible using KMinion.
- Kafka versions: Supports all Kafka versions v0.11+
- Supported SASL mechanisms: plain, scram-sha-256/512, gssapi/kerberos, oauthbearer
- TLS support: TLS is supported, regardless whether you need mTLS, a custom CA, encrypted keys or just the trusted root certs
- Consumer Group Lags: Number of messages a consumer group is lagging behind the latest offset. Group offsets are
read either via the Admin API (default) or by consuming the
__consumer_offsetstopic, which additionally exports the number of offset commits per group - Log dir sizes: Metric for log dir sizes either grouped by broker or by topic
- Broker info: Metric for each broker with its address, broker id, controller and rack id
- Configurable granularity: Export metrics (e.g. consumer group lags) either per partition or per topic. Helps to reduce the number of exported metric series.
- End to End Monitoring: Sends messages to its own topic and consumes them, measuring a messages real-world "roundtrip" latency. Also provides ack-latency and offset-commit-latency. More Info
- Connection health check: Optionally opens a fresh connection to every broker periodically, to detect a broker that has silently stopped accepting new Kafka connections
- Configurable targets: You can configure what topics or groups you'd like to export using regex expressions
- Multiple config parsers: It's possible to configure KMinion using YAML, Environment variables or a mix of both
You can find a list of all exported metrics here: /docs/metrics.md
All images will be built on each push to master or for every new release. You can find an overview of all available tags in our DockerHub repository. Release images are also published to Amazon ECR Public.
docker pull redpandadata/kminion:latestPrebuilt binaries for Linux, macOS and Windows (amd64 and arm64) are attached to every GitHub release.
To build KMinion from source you need the Go version declared in go.mod or newer:
go build -o kminion .A Helm chart will be maintained as part of Redpanda's helm-charts repository.
All options in KMinion can be configured via YAML or environment variables. Configuring some options via YAML and some via environment variables is also possible. Environment variables take precedence in this case. You can find the reference config with additional documentation in /docs/reference-config.yaml.
If you want to use a YAML config file, specify the path to the config file by setting the env variable
CONFIG_FILEPATH.
KMinion listens on port 8080 by default (exporter.port) and serves:
/metrics: the Prometheus metrics/ready: withminion.consumerGroups.scrapeMode: offsetsTopic, returns 503 until KMinion has caught up with the__consumer_offsetstopic, and 200 after that. In the defaultadminApimode it always returns 200. It does not check the connection to Kafka
There are three separate Grafana dashboards that can be used as inspiration in order to create your own dashboards. Please take note that these dashboards might not immediately work for you due to different labeling in your Prometheus config. They may also not cover newer metrics, such as the connection health check metrics.
Cluster Dashboard: https://grafana.com/grafana/dashboards/14012
Consumer Group Dashboard: https://grafana.com/grafana/dashboards/14014
Topic Dashboard: https://grafana.com/grafana/dashboards/14013
This repo contains a docker-compose file that you can run on your machine. It will spin up a single-node Redpanda cluster and starts KMinion on port 8080 which is exposed to your host machine:
# 1. Clone this repo
# 2. Browse to the repo's root directory and run:
docker compose upWe use Slack to communicate. If you are looking for more interactive discussions or support, you are invited to join our Slack server: https://redpanda.com/slack
KMinion is distributed under the MIT License.