A simple service collecting system metrics like CPU, memory, disk space and network usage and pushing them to a Prometheus instance using the write API. Specifically, this is useful if you want to send simple system metrics from servers, which are not or shall not be accessible via a web API.
Note
This service has a very limited set of metrics used for very simple system monitoring. If you are looking for a more comprehensive solution, please take a look at the following projects.
On Debian based systems, you can download the Debian package from the Releases Page for your system and install it using dpkg.
dpkg -i sysinfo-exporter-*.deb
After that, open the configuration file placed in /etc/system-exporter/config.toml and enter the Prometheus endpoint and other desired configuration parameters.
Simply download the latest binary form the Releases Page for your system architecture and install it on your system.
After that, choose a way for configuring the service.
The configuration can either be passed via endironment variables (prefixed with SYSEXP_) or via a TOML config file passed by the --config parameter.
Here you can find an example configuration.
# The write API endpoint of the target prometheus instance.
endpoint = "https://prometheus.example.com/api/v1/write"
# When the target prometheus instance requires authentication,
# here you can specify the type and credentials.
# [auth.basic]
# username = ""
# password = ""
# [auth.bearer]
# token = ""
# The log level of the application logger.
# Besides the global log level, you can specify levels for filters.
# Here you can find more information:
# https://docs.rs/tracing-subscriber/latest/tracing_subscriber/filter/struct.EnvFilter.html#directives
loglevel = "info"
# The interval (in seconds) in which system information data is
# refreshed and pushed to the prometheus instance.
interval_seconds = 15
# You can specify additional labels, which are added to each
# metric sent from this instance.
[labels]
location = "home-1"