Using the AI Assistant to examine RUM data
The AI Assistant in Splunk Observability Cloud can help you get faster insights from your Splunk Real User Monitoring (RUM) data - but like any AI tool, the quality of your results depends on how you frame your questions. This article shares practical tips for writing effective prompts, along with example prompts organized by topic to help you get started.
Tips for writing effective prompts
Keep these guidelines in mind when crafting your prompts:
- Be specific. Include your app name, environment name, and any other relevant context. The more precise your prompt, the more useful the response.
- Use your current filters. The AI Assistant is aware of the page-level filters currently applied in your session, so factor those into how you phrase your query.
- Iterate. If an initial response isn't quite what you need, try rephrasing or adding more detail.
Example prompts by topic
Overall health and performance
Use these prompts to get a broad view of how your application is performing:
- What are the problems with my <app_name>?
- What are the problems with my latest app version for <app_name>?
- Get application performance metrics, page load time, network request latency, and Core Web Vitals metrics.
- Get the app startup times for <app_name>.
- Show the impacted sessions or spans for <ErrorID>.
- Show me the <tag_name01> and <tag_name02> for the metric <metric_name>.
- Show me top crashes for <app_name>.
- Show me the web vitals for <app_name>.
- Give me a summary of all crashes for <app_name>.
- What are the network latency issues with my <app_name>?
- Show me the top 5 slowest pages for <app_name>.
Alerts
Use these prompts to review alert status and health violations:
- Are there any active alerts for <app_name>?
- Summarize health violations for <app_name>.
- Show me the detectors set for my <app_name>.
What's next
After reviewing the AI Assistant's responses, use the thumbs up or thumbs down controls to rate the quality of the answers - your feedback helps improve future results.
If you have additional example prompts or suggestions for this guidance, use the feedback option below to share them.
In addition, these resources might help you understand and implement this guidance:
- Splunk Lantern Article: Implementing key use cases for the Splunk AI Assistant for SPL
- Splunk Lantern Article: Enabling the Splunk Enterprise AI Assistant for SPL