Investigating a sudden performance degradation by using the Adaptive Data Collector
By using the Adaptive Data Collector, you can leverage adaptive, context-aware data collection to gain relevant diagnostic data without relying on continuous high-volume data collection.
Scenario
Apex Global operates Apex-Retail, an online service for selling pharmaceutical products. Apex-Retail is a customer-facing application with variable traffic patterns and periodic changes caused by traffic spikes. Under normal conditions, the service operates within predictable performance thresholds.
Sarah, a tenant administrator at Apex Global, is responsible for monitoring the Apex-Retail service health and responding to performance-related incidents. Sarah observes that the Apex-Retail service is responding slowly. The performance degradation is recent and is not related to an ongoing maintenance window. To quickly mitigate the situation, Sarah needs to understand which component of the Apex-Retail service is experiencing an issue. It could be one of the following components:
- The web server, which handles the front end of the service
- The application server, which handles the back end of the service
- The database server, which handles storage for the service
Sarah does not want to waste time collecting data for all the available metrics. She wants to collect data for only the affected metrics.
To diagnose the issue, Sarah runs the Adaptive Data Collector from BMC Helix Ops Swarmer. The Adaptive Data Collector takes the following actions:
- Identifies a disk-related issue on the database server
- Increases data collection on the following metrics:
- KPI Working set (KB)
- Page faults per second
- Page file bytes (KB)
- Private memory (KB)
- Process status
- Does not collect data on the remaining metrics
With an increased data collection on specific metrics, Sarah can perform the following actions:
- Analyze the collected metrics to correctly understand the issue
- Take corrective action faster to resolve the issue
By using the Adaptive Data Collector, Sarah could achieve the following results:
- Avoid high-frequency polling for data collection and save on operational costs
- Quickly identify the area of concern and collect metrics around it
- Reduce the response time because of focused data collection
The following image displays the metric collection before running the Adaptive Data Collector, where all metrics are enabled:
The following image displays the metric collection after running the Adaptive Data Collector, where only specific metrics are enabled:
Workflow for investigating performance degradation by using the Adaptive Data Collector
| Task | Action | Product | Role | Reference |
|---|---|---|---|---|
| 1 | Run the Adaptive Data Collector by using Microsoft Teams or BMC HelixGPT. | Microsoft Teams or BMC HelixGPT | Administrator | Agentic AI capabilities in BMC Helix Operations Management |
| 2 | Analyze the collected metrics to take corrective action. | BMC Helix Operations Management | Administrator | Viewing collected data |
To run the Adaptive Data Collector
- Open BMC HelixGPT.
- Run a prompt to trigger the Adaptive Data Collector.
For example:
Enable context-aware data collection on the Apex-Retail service.
Instead of BMC HelixGPT, you can use Microsoft Teams to run the prompt.
Related topics