Monitoring data processed
View the volume of data that is processed by Cloud Pak for AIOps to see whether the OpenShift Container Platform cluster is sized properly.
Use the following topics to view the volume of data that is processed by Cloud Pak for AIOps.
Events
For a rough estimate of the number of events, go to the AIOps Insights dashboard. The AIOps Insights dashboard can be accessed from the administration page of IBM Cloud Pak for AIOps, under Overview > Quick navigation > AIOps insights. Alternatively, click the navigation icon (four horizontal bars) to go to the main navigation menu. Then, click Operate > AIOps insights.
The noise reduction graph shows the number of events over the selected time period, such as 24 hours, 7 days, or 30 days. You can use the average of the events when you use the Cloud Pak for AIOps deployment sizing.
For more information about precise distribution of events, see Enabling monitoring for user-defined projects.
When the OpenShift user workload monitoring is enabled, you can view the metrics in the OpenShift dashboard.
The following Prometheus query can be used to view the number of events, which can be added in the Cloud Pak for AIOps console console under Observe > Metrics:
sum(irate((flink_taskmanager_job_task_operator_KafkaSourceReader_topic_partition_currentOffset{topic="cp4waiops_cartridge_lifecycle_input_events"}>0)[1m:30s]))
Topology
To view the number of resources, click Resource Management, and view the resource count. For example, you can see 864 resources in the following example.
For more information about the Resource management, see The Resource management page.
Logs
To view the number of logs, extract the count from the Elastic backend:
-
Log in to your OpenShift Container Platform cluster by using the command:
oc login <server> -u <cluster_username> -p <cluster_pass>
Where
<server>
is the hostname with port number, such as,https://localhost:8443
.<cluster_username>
is the username.<cluster_pass>
is the password.
-
When you are logged in to the cluster, use the following command in your command-line:
while true; do oc port-forward $(oc get po | grep iaf-system-elasticsearch-es-aiops-0 | awk '{print $1}') 9200; done
-
On a separate command-line window, run the following set of commands:
export username=$(oc get secret iaf-system-elasticsearch-es-default-user -o jsonpath="{.data.username}"|base64 -d); export password=$(oc get secret iaf-system-elasticsearch-es-default-user -o jsonpath="{.data.password}"|base64 -d); curl -u $username:$password -XGET https://localhost:9200/_cat/indices --insecure | sort | grep logtrain curl -u $username:$password -XGET https://localhost:9200/$INDEX/_count --insecure
The outputs of the preceding curl commands provide the daily count of the logs processed. To align with the Cloud Pak for AIOps deployment sizing, you can average the daily count of logs, or pick the maximum value from both of the curl commands.
Metrics
To view the number of Metric KPIs being tracked, click AI Model Management > Metric anomaly detection.
The number of KPIs is noted under the Metrics analyzed
field. For example, 146 metrics were analyzed in the following example.
For more information about Metric anomaly detection, see the following links: