Closed fjtirado closed 2 months ago
@fjtirado IMO we should record also the number this node has been executed. The reason is because you can calculate the mean (duration) / hits = average so we have an idea of the duration of the node.... min max would be interesting as well.
@fjtirado IMO we should record also the number this node has been executed. The reason is because you can calculate the mean (duration) / hits = average so we have an idea of the duration of the node.... min max would be interesting as well.
@elguardian I think this is automatically done by the Distribution summary micrometer. This is an example of what is sent to the elastic search for the proccess instances distribution summary
{"@timestamp":"2024-04-22T11:07:47.845Z","name":"kogito_process_instance_duration_seconds","type":"distribution_summary","app_id":"default-process-monitoring-listener","artifactId":"serverless-workflow-expression-quarkus","process_id":"division","version":"1.0-SNAPSHOT","count":1,"sum":0.061,"mean":0.061,"max":0.061}
Fixes https://github.com/apache/incubator-kie-kogito-runtimes/issues/3483 Add "kogito_node_instance_duration_milliseconds" distribution summary. Different to the other two summaries, times are collected in milliseconds rather than seconds