Real-time metrics

Details about the real-time metrics (active user count etc) provided by Butler.

Real time data on active user sessions

Butler listens to session start/end messages from Sense’s log4net logging framework, and keeps records of both how many and which users currently have active sessions.
Ideal would be if the Sense APIs would provide a list of currently active users, that us however currently not available. For that reason Butler’s tactic is to listen to log events to determine when users start and end sessions.

There is however a downside to this approach.

When Butler is first started, it will not know what sessions are currently active. Instead, it will collect that inforamation as time passes.
This means that the active user count, as well as what specific users are active, might be incorrect when Butler is first started.
The data will then become more and more exact as users end their sessions, and later restart them (Butler will then capture the session restart, and update its data structures accordingly).

If there are demands for enterprise grade monitoring of a Qlik Sense environment, the Butler SOS tool focuses exclusively on this and might be worth looking into.

Active users data as MQTT messages

Butler publishes a set of MQTT messages relating to active sessions.
Whenever a session starts or ends, two MQT messages will be sent:

  • The new number for number of active user sessions is posted to the MQTT topic defined by the config entry Butler.mqttConfig.activeUserCountTopic
  • A string array with usernames of all users with active sessions is posted to MQTT topic defined by Butler.mqttConfig.activeUsersTopic

Making this data available as MQTT messages, it can then be used in real-time dashboards or for alerting purposes.
For example, using Node-RED’s dashboard module, it is trivial to create a real-time updating dashboard like this one (showing data for the last few days):

alt text



Last modified 2021-04-22: First commit (c105eb7)