Monitoring Flussonic with Prometheus

You can monitor your Flussonic server(s) and collect the data with the help of Prometheus. In recent versions server monitoring was performed through SNMP, parsing JSON files and Pulse tab in Flussonic UI.

What is Prometheus?

Prometheus is an open-source systems monitoring and alerting toolkit. It stores the data in a time series database, built specifically for handling metrics and events or time-stamped measurements. It contains the data in the form of "key (metric's name) - value" pairs for each point in time. This way, it is accessed effectively.

For data visualization Prometheus is integrated with Grafana.

Why Prometheus?

quick and efficient data access due to time series database;
alerts via methods such as email, on-call notification systems, and chat platforms (Slack, Telegram, etc.) so that you can choose the most suitable notification sending method for yourself;
displays data of a group of servers on one dashboard, which is convenient for viewing the status of multiple servers at a time;
stores data for some time, which allows you to monitor the state of the server or a group of servers over time;
monitors not only the whole server but also its separate streams.

Getting data from Flussonic

Here we will give you some instructions on how to access the metric values by Flussonic.
There are only a few simple steps:

Step 1:
First, you should download Prometheus for your platform and install it. This guide will be of great use for you.

Step 2:
Configure Prometheus through the prometheus.yml file adding these lines to the basic configuration:


scrape_configs:
  - job_name: 'myflussonic'
    metrics_path: /flussonic/api/metrics
    basic_auth:
      username: 'flussonic'
      password: 'password'
    static_configs:
      - targets: ['FLUSSONIC-IP:80']

Parameters:

job_name — the name of the process;
matrics_path — path to the metrics calculated by Flussonic;
username — your username as in Flussonic user account;
password — your password as in Flussonic user account;
targets — list of server URLs that you want to monitor.

Step 3:
Add a dashboard to Grafana, using this link to visualize your data.

Metrics to monitor

Here are metrics for monitoring both the whole server and separate streams.

Server metrics

Metrics	Unit	Description
flussonic_version_info	-	Current Flussonic version
flussonic_boot_time_seconds	s (in Unixtime)	Flussonic boot time
flussonic_memory_usage	%	Memory usage
flussonic_clients	-	Clients number
flussonic_streams	-	Streams number
flussonic_opened_files	-	Opened files number
flussonic_scheduler_load	-	Erlang scheduler
flussonic_cpu_usage	%	CPU usage
flussonic_disk_size	bytes	Disk size
flussonic_disk_usage	%	Disk usage
flussonic_disk_io_usage	%	Disk I/O usage
flussonic_net_input_bytes	bytes	Network interface input
flussonic_net_output_bytes	bytes	Network interface output
flussonic_dvb_card_strength	-	DVB card signal stength
flussonic_gpu_dec	%	GPU decoder usage
flussonic_gpu_enc	%	GPU encoder usage
flussonic_gpu_temp	°C (degree Celsius)	GPU temperature

Stream metrics

Metrics	Unit	Description
flussonic_stream_client_delay	ms	Client delay
flussonic_stream_bytes_in	bytes	Number of input bytes
flussonic_stream_bytes_out	bytes	Total number of bytes in the output. Calculated as a number of packets multiplied by 188.
flussonic_stream_bitrate	kB/s	Bitrate
flussonic_stream_start_running_at	ms	Start time
flussonic_stream_retry_count	-	Number of automatic retries
flussonic_stream_client_count	-	Number of clients
flussonic_stream_dvr_info_disk_size	bytes	DVR disk size
flussonic_stream_dvr_info_depth	s	DVR depth
flussonic_stream_dvr_info_duration	s	DVR duration
flussonic_cbr_mpegts_trimmed	bytes	Number of trimmed data. Calculated for MPEG-TS muxer.
flussonic_bytes_out	bytes	Number of output bytes. Calculated for payload, stuffing and fillers separately.

Now you can monitor your Flussonic server or a group of servers as well as streams with Prometheus and Grafana.