Question 850
A business application becomes unreachable while all monitored probes look normal, and the team concludes the failure is impossible because the analytics page shows no anomaly. Once the issue is resolved, which ongoing practice BEST prevents this type of failure?
Explanation
Correct Answer
A lasting control should detect the original error before release. Document which IBA probes support each service SLO and where monitoring coverage has gaps. That is important in this case because probe anomalies represent observations of defined metrics, while service health must be established from the relevant intent dependencies and traffic results. The scenario was: A business application becomes unreachable while all monitored probes look normal, and the team concludes the failure is impossible because the analytics page shows no anomaly.
Incorrect Answers
A valid analytics conclusion depends on receiving timely samples from the selected device and monitoring pipeline, not simply on an existing probe definition. For the current prevention question, the evidence instead states: A business application becomes unreachable while all monitored probes look normal, and the team concludes the failure is impossible because the analytics page shows no anomaly. Here, probe anomalies represent observations of defined metrics, while service health must be established from the relevant intent dependencies and traffic results.
Pair every analytics-triggered incident with recorded before-and-after probe and service evidence. That concerns verifying remediation against IBA evidence, while the issue here is IBA probe anomaly versus affected service. Correlate the specific probe and its target objects with service-level evidence, device state and path tests before assigning incident scope.
An Apstra IBA probe defines a supported analytical measurement or condition over selected network resources and requires appropriate source data. However, the present prevention decision is grounded in: A business application becomes unreachable while all monitored probes look normal, and the team concludes the failure is impossible because the analytics page shows no anomaly. For this issue, correlate the specific probe and its target objects with service-level evidence, device state and path tests before assigning incident scope.