• This plugin will post alarms and metrics to a Kafka node.
• Different topics for alarms and metrics can be created to distribute the events and to facilitate the integration with third party consummers.
• The Kafka plugin will flush received events every 30 seconds to the Kafka node.
• If the node cannot be reached, events will be queued until connectivity is back or until queue is full.

• Press Create or select Edit in the table action menu to open the Kafka plugin editor
• Set Kafka plugin parameter
  • Active
  • Title
  • Bootstrap servers
  • Client Id
  • Alarms topic
  • Metrics topic
  • User
  • Password
  • Properties
  • Max queue size
  • Send alarms
  • Send metrics

Note: If the alarm or metric queue is full, incoming events will be discarded until space is made in the queue.

You can use variable names in topic to dynamically redirect the events into specific topics. You can use metric or alarm tag property names as variables:

Example: Let's consider following metric:

{"timestamp":1594742051,"value":4.0,"key":"sap.hana.HANA_THREADS.THREADS_COUNT","tags":{"instance":"aglgesrvhana02b","sid":"HA5","group":"Agentil"}}

If you set “metrics-%group%-%sid%“ in metric topic, this metric will be put in topic: metrics-Agentil-HA5

• Properties are key/value pairs separated by semicolon: val1=12;val2=no
• The list of exhaustive properties can be found here
• Default settings:
  • Batch size is set to 500KB
  • Linger time is set to 10ms (Time to wait for new events before sending an incomplete batch)
  • Compression is OFF (see property compression.type=gzip)