EmbySessionLogging

Emby server session logging will allow you to monitor what is happening on your Emby server. This can help detect load issues and overall server health. Once you have session logging set up you can view the data in metrics monitoring tool like Grafana.

Run the logging script

• Download and install the latest version of Python
  Windows Python distribution Miniconda - https://docs.conda.io/projects/miniconda/en/latest/
• Download the Python script emby_session_logging.py and save it locally
• In the Emby web client dashboard go to the API Keys area, create a new API key
• Use the following command to start logging, replacing <api_key> with the API key you created:
  python emby_session_logging.py <api_key>
• The script logs session data into a SQlite DB file sessions.db in the same dir as the script is run
• The script will run logging the Emby session info every 60 seconds

Check logging

• Use the following command to dump all the logged data:
  python emby_session_logging.py dump
• This will dump all the logged data to the console

Setup Grafana

• Download and install the latest version of Grafana:
  https://grafana.com/grafana/download
• Install the SQLite datasource plugin:
  https://grafana.com/grafana/plugins/frser-sqlite-datasource/
• Start up and verify Grafana is working and you can see the interface:
  http://localhost:3000

Setup datasource

• In Grafana create a new data source connection in Grafana
• Click the 3 bars, top left, to bring up the menu
• Click Connections and then click data sources, in the top right click "+ Add new data source"
• Enter SQlite in the search box
• You should see the SQlite data source plugin (if you dont then it is not install correctly go back and work out why)
• Click the SQlite data source plugin
• Give your data source a name in the name field, something like "emby_sessions"
• You need to fill in the path to the sessions.db file created when running the logging script

Grafana dashboard

• Create a new dashboard
• Add a visualization to the new dashboard
• In the data source drop down select the SQlite data source create above
• Copy and paste the following query into the query area
• Click Apply, you now have a new chart with the logged session data

SELECT 
	event_date as ts, 
	CAST(json_extract(data, '$.as') AS FLOAT) as "Active",
	CAST(json_extract(data, '$.p') AS FLOAT) as "Playing",
	CAST(json_extract(data, '$.vt') AS FLOAT) as "VTrans",
	CAST(json_extract(data, '$.at') AS FLOAT) as "ATrans",
	CAST(json_extract(data, '$.hvd') AS FLOAT) as "HwDecode",
	CAST(json_extract(data, '$.hve') AS FLOAT) as "HwEncode"	
FROM session_log
WHERE event_date BETWEEN ${__from:date:seconds} AND ${__to:date:seconds}
GROUP BY 1 ORDER BY 1 ASC