Auto Analysis

Kloudfuse uses the Bullseye service for analyzing signals, and correlating them with signals within the same stream or across streams. This analysis contributes to the reduction in the effort necessary to debug issues and reduce problematic areas, in minutes.

You can configure Bullseye to run an analysis when an alert fires, or on an ad-hoc basis. Bullseye is configured to execute all alerts and signals.

Ad-Hoc Analysis

  1. In the APM Services interface, or another interface that shows services monitored by Kloudfuse, chose the service you plan to analyze.

    Here, we plan to analyze the frontend service; click the name of the service:

    Analyze frontend service

  2. The Service Detail interface appears.

    If you enabled alerts and ASM for the service, the interface displays a green signal and a number that represents monitoring assignments/alerts. Here, there are 3, and status for all is OK.

    Analyze frontend service

  3. Click ASM button to see the alerts.

  4. From the list of options, select the alert you plan to analyze.

    Here, we plan to examine the alert "anomaly - 90th percentile request latency".

    examine the anomaly alert

  5. The Alert Detail interface appears.

  6. At the bottom of the interface, find an Event to examine, and click its Analyze (Analyze) button.

    start analysis

  7. The Alert Event Analysis interface appears, as a service map view.

  8. In the service map, find the service with which to correlate.

    Here, let us examine recommendationservice.

    Examine the service for correlation

  9. Click the node (circle) that represents the service you plan to correlate with your initial service.

    Choose node/service for correlation

  10. In the Enable Correlation widget, choose what metrics you plan to correlate.

    Here, we plan to correlate both RED Metrics and USE Metrics.

    Choose both RED and USE metrics to correlate

  11. Notice that both correlation types appear under the Selected Correlation Service on the right side of the Alert Event Analysis interface.

  12. Select other services for correlation, repeating steps 9 through 11 as many times as necessary.

    In this example, we selected reommendationservice downstream of the target node, and frontendproxy upstream of the target.

  13. In the Selected Correlation Service section, click the Run Correlation (Run Correlation) button.

    auto analyze run correlation

  14. After the correlation runs, examine the results.

  15. Any correlated metrics appear in the result information, under the Metrics tab:

    result metrics

  16. The Logs tab shows the list of all relevant logs.

    result logs

  17. The Traces tab shows the list of all relevant traces.

    result metrics