[FEAT] Add OpenMetrics / Prometheus metrics endpoint

[kinghrothgar]

It would be very nice to be able to pull data from scrutiny into Prometheus. This would allow me to leverage my standardized alert system.

https://prometheus.io/docs/instrumenting/writing_exporters/

It should feasibly be a /metrics route that translates the results of a DB query into the simple OpenMetrics format. If there is interest in merging this feature, I can look into getting together a PR.

[AnalogJ]

Hey, so the idea would be to allow Prometheus to query smart data from Scrutiny?
Doesn't Prometheus have a SMART data collector already?

[kinghrothgar]

There seems to a not particularly maintained exporter here (last interaction of the maintainer was over a year ago): https://github.com/Sheridan/smartctl_exporter

The only other option I have found is to do something like this with a node_exporter script, but this example script leaves much interpretation to be desired (as noted in the top comment): https://github.com/prometheus-community/node-exporter-textfile-collector-scripts/blob/master/smartmon.sh

[dvonessen]

Hey! Is there any work on this feature? Or is this abandoned?

[Tin-Joy59]

Hey, I've been looking for the same thing! I've been using Scrutiny for a very long time and would like to bank on its reliability to push stats in Prometheus and Grafana.

[AnalogJ]

I'm open to contributions to support this kind of functionality, however I dont use Prometheus so I will be unable to develop this myself unfortunately

[stavros-k]

Bringing this feature will benefit Windows users as well, as there are no exporters for that on Windows (that I know of).
Meaning that scrutiny will be the only(?) maintained exporter for SMART data in that platform

[proffalken]

Whilst I don't have the skills to develop this, I'm more than happy to test it if it gets developed.

[christianhorn]

I think this could be realized in running PCP's pmcd, pmda-smart (to read smart data) and pmproxy which can offer the data in Openmetrics format.

[MomentumTender7962]

So, I might have figured out a workaround to this. You can connect Grafana to the InfluxDB that Scrutiny uses to store all the smart data.

However, I have not found out a way to find the credential information scrutiny uses to initialize the InfluxDB. I did this myself using Docker Compose environment variables:

environment:
- DOCKER_INFLUXDB_INIT_MODE=setup # MUST COMMENT OUT AFTER FIRST LAUNCH. It will setup the database correctly but will not connect correctly until relaunched in not setup mode
- DOCKER_INFLUXDB_INIT_USERNAME=admin
- DOCKER_INFLUXDB_INIT_PASSWORD=SomePassword
- DOCKER_INFLUXDB_INIT_ORG=homelab
- DOCKER_INFLUXDB_INIT_BUCKET=scrutiny
- DOCKER_INFLUXDB_INIT_ADMIN_TOKEN=SomeToken

After that edit the influxdb settings in the scrutiny.yaml file:

influxdb:
scheme: 'http'
host: 0.0.0.0
port: 8086
token: SomeToken
org: 'homelab'
bucket: 'scrutiny'
retention_policy: true

Make sure the values for token, org and bucket match in both the docker compose and yaml files.

After that, you can add the the influxdb as a data source by selecting Connections > Data Sources > Add New > InfluxDB.

• Set the url to the url and port of the influxdb instance
• Select the 'Query Language' as 'Flux'
• enter the org, bucket and token information above

By default, the influxdb is only exposed to scrutiny in the docker compose stack. You'll have need to make sure it is also exposed to a network that grafana is on. For example, in my compose I have:

networks:
  - default
  - grafana_net

And then at the end declare the networks

networks:
grafana_net:
external: true

I am still working out how to query the information since I have never queried data from an InfluxDB before. But this is how I connected it, and have been able to list the buckets and some fields and tags. But perhaps someone smarter and more familiar with InfluxDB can add to this.

[jianyun8023]

I'm a fan of Scrutiny and wanted to export its metrics to Prometheus. Since I couldn't find an existing solution, I built my own exporter.

I've also packaged it into a container image and created a Grafana dashboard to get started quickly. If you're interested, you can check out the repository here:

• https://github.com/jianyun8023/scrutiny-exporter

Why not use InfluxDB? Because the data in InfluxDB lacks information that helps identify, such as host and device names.

[jianyun8023]

My OCD tendencies kicked in, and after reviewing the project code, I decided to contribute.

I created two PRs to support this feature—one for a bug fix and one for the feature itself.
I used AI to assist with programming, designed the implementation plan, and reviewed and validated the functionality. If you have any concerns, feel free to close the relevant PRs.

• Fix smart attributes query to use limit() instead of tail() for fetching recent entries #829
• Implement Prometheus metrics support in Scrutiny #830