This release is full of exciting things.
Prometheus support has been added, and that makes this application dual purposed: you may store your metrics in InfluxDB or Prometheus. The same dashboards are available for both data sources!
To enable Prometheus support, set
prometheus in your up.conf config file. Then add a job definition to your prometheus config. The poller runs on port 9130 by default, so it looks like this:
- job_name: 'unifi' static_configs: - targets: ['127.0.0.1:9130']
The documentation will be updated eventually, but keep in mind you need port 9130 available if you run this in Docker.
Prometheus isn’t the only exciting news! The poller has been shoving lots of useless data into your influx db, and this release fixes that. All of the devices have had their tags unified and cleaned up tremendously. Many unused and unneeded metrics were removed from the output. These removed metrics include those for
down USW ports. That means it may look like some ports are missing - this is expected. This is a pretty big cleanup, and should save on storage and memory usage.
The UAP VAP data was still corrupted. It’s been messed up since this thing started, and I finally spent enough time with the data (while adding Prometheus support) to figure it out. I ported the fixes to the influx code. This creates a new table, and many of the UAP graphs had to be updated. This results in more accurate data presentation.
All five dashboards received updates. When you upgrade the poller, upgrade your dashboards.
As much as it pains me, I also recommend deleting your influx database and starting from scratch. This will improve performance and your experience with the updated dashboards. It’s simple:
$ influx > drop database unifi > create database unifi
If you’re running influx (or prometheus) on a small system like raspberry pi you probably run into memory pressure. A new option to disable site data storage has been added to the poller. This removes the data shown on the Network Sites dashboards, and will save a little bit of data space and memory consumption.
Maintaining this is a lot of work…. Thank you to @SuperQ for the help making this possible! Please let me know if you run into any issues with the new dashboards!