Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/sabeechen/rabbitmq_exporter
A simple rabbitmq prometheus exporter
https://github.com/sabeechen/rabbitmq_exporter
Last synced: 15 days ago
JSON representation
A simple rabbitmq prometheus exporter
- Host: GitHub
- URL: https://github.com/sabeechen/rabbitmq_exporter
- Owner: sabeechen
- License: unlicense
- Created: 2024-10-13T16:37:07.000Z (3 months ago)
- Default Branch: master
- Last Pushed: 2024-11-18T21:05:04.000Z (about 2 months ago)
- Last Synced: 2024-12-17T14:12:53.460Z (17 days ago)
- Language: Python
- Size: 12.7 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# A Very "Basic" RabbitMQ Exporter
Provides a "super simple" RabbitMq exporter that exposes metrics and healthz endpoints. Only a handful of global celery queue metrics are reported:
- `rabbitmq_queue_messages` - The message queue length
- `rabbitmq_queue_message_bytes` - The message queue size in bytes
- `rabbitmq_queue_messages_ready` - The number of messages ready to be consumed
- `rabbitmq_queue_messages_unacknowledged` - The number of messages that have been delivered but not yet acknowledged
- `rabbitmq_queue_consumers` - The number of consumers on the queue
Each metric exposes a 'queue' label with the queue name.
### Endpoints
- `/metrics` - The metrics endpoint
- `/healthz` - The healthz endpoint. Returns a non-200 response when it can't reach RabbitMQ successfuly.
Endpoints are hard-coded to listen on port 9090.
# Why?
RabbitMQ exposes a native prometheus management plugin. If you're able to use that its much better, but services like AmazonMQ don't expose that
endpoint/plugin becaus they'd rather you pay a lot of money to deal with their eccentric CloudWatch API. This is a burden if you're just running a
simple prometheus stack, so this little script exists to export some simple information from RabbitMQ's management API directly. Its designed to
work well in a prometheus/kubernetes environment.
This exporter is simple, stateless, low resource, and reliable. The server is implemented using pure async methods. Each request to the `/metrics` triggers its own http request to RabbitMQ to get the queue information in real-time (eg no side-polling).
# Usage
### Run it
You can call the script directly with a `--url` encoding the user/pass/host:
```bash
python exporter.py --url amqps://user:[email protected]
```
The scheme/port/path for the url is always ignored and https:443 is used against the specified hostname, pass `--verify-ssl=false` to disable SSL verification. The user and pass can also be specified using `--user` and `--password`.
### Run it in Docker
```bash
docker run -p 9090:9090 -e MQ_URL=https://user:[email protected] ghcr.io/sabeechen/rabbitmq_exporter
```
The args can also be specified as environment variables:
- `MQ_URL` - The RabbitMQ API endpoint (amqp:// or https:// with or without credentials)
- `MQ_USER` - RabbitMQ username
- `MQ_PASSWORD` - RabbitMQ password
- `MQ_TIMEOUT` - RabbitMQ http timeout (seconds)
- `MQ_VERIFY_SSL` - Verify SSL certs from the broker (true/false)