Queue depth and consumer monitoring

Options
Suganya
Suganya Member Posts: 1

1.To identify the queue which is having current queue depth greater than (specific threshold e.g 1000)
2.To identify the queue which is having current queue depth greater than (specific threshold e.g 1000) and also consumer is not connected.

Which will be the effective way to monitor from support perspective.
1) to write a job to run specific times to see any clients breached
2) any other tools which will send notification to team.
3) Prometheus

Tagged:

Comments

  • marc
    marc Member, Administrator, Moderator, Employee Posts: 923 admin
    Options

    Hi Suganya,
    The best way to be alerted when queue depth reaches a certain point is to configure alert thresholds on an endpoint. Alert thresholds can be set by message size (% of queue quota) or by the number of messages and also allow for a "clear" threshold to be set so you can also be alerted when queue depth goes back to a nominal state.
    One way you can set the alerts is via the PS+ Manager as seen here:

    When the threshold is passed a VPN_AD_MSG_SPOOL_HIGH" alert would be triggered. You can subscribe to these alerts via syslog or over the message bus itself. More info on monitoring options here: https://docs.solace.com/Best-Practices/Monitoring-BP.htm#mc-main-content

    As far as your second question you could also monitor syslog events to see when clients connect/disconnect or you can use SEMP (our RESTful management API) to query and check. More info on SEMP here

    From a monitoring perspective we do offer the PubSub+ Monitor tool, but we have also seen customers use all sorts of different tool sets to monitor our brokers so it really differs on a user by user basis. With the ability to receive alerts via syslog or over the message bus and also having SEMP you can really integrate with most any monitoring system.