https://github.com/remram44/fs-latency-exporter

Measures read latency from a locally-mounted filesystem, exposes as metric
https://github.com/remram44/fs-latency-exporter

metrics prometheus-exporter

Last synced: 5 days ago
JSON representation

Measures read latency from a locally-mounted filesystem, exposes as metric

• Host: GitHub
• URL: https://github.com/remram44/fs-latency-exporter
• Owner: remram44
• License: mit
• Created: 2024-10-23T02:31:18.000Z (22 days ago)
• Default Branch: trunk
• Last Pushed: 2024-11-06T18:33:18.000Z (8 days ago)
• Last Synced: 2024-11-06T19:35:50.763Z (8 days ago)
• Topics: metrics, prometheus-exporter
• Language: Rust
• Homepage:
• Size: 52.7 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.txt

Awesome Lists containing this project

README

        
Filesystem Latency Exporter

===========================

This is an exporter for Prometheus, measuring the latency of random read to a local file.

It can be used with networked file systems as well, if they are mounted on the local machine.

Reads use direct I/O when available (UNIX), picking a random 4096-byte block in the file.

![Screenshot of a Grafana heatmap panel](screenshot-grafana.png)

Example usage:

```

# Create a large file on the target filesystem

dd if=/dev/urandom of=/mnt/100GB.bin bs=10M count=10000 status=progress

# Run the exporter

./fs-latency-exporter --metrics 127.0.0.1:8080 /mnt/100GB.bin

# Grab the metrics (in another terminal, or via Prometheus)

curl -s http://127.0.0.1:8080/metrics

```

The exposed metrics are:

- `errors_total`, a counter of errors encountered when reading and seeking

- `read_time_seconds`, a histogram for the duration of the random reads