Documentationbreadcrumb arrow Grafana Cloudbreadcrumb arrow Adaptive Telemetrybreadcrumb arrow Adaptive Profilesbreadcrumb arrow Guidesbreadcrumb arrow Troubleshoot Adaptive Profiles insights
Grafana Cloud

Troubleshoot Adaptive Profiles insights

Insights are generated after a service completes a boost and sampling returns to normal.

This guide helps diagnose missing or delayed insights.

Issue 1: Insights do not appear

Insights are not generated unless certain conditions are met. They require:

  • The Grafana LLM plugin to be enabled (optional feature)
  • At least one completed boost (manual or automatic)
  • Enough high-resolution data captured during the boost

If any of these conditions are missing, insights do not appear.

Note: The LLM plugin is optional. Adaptive Profiles works with sampling capabilities even without the LLM plugin enabled, but AI-powered insights won’t be generated.

What you can do

When insights are missing:

  • Trigger manual boosts for one or more services
  • Wait until sampling returns to baseline

Insight generation cannot be forced, and timing varies depending on:

  • Boost frequency
  • Available profiling data
  • Backend processing load

Insights should eventually appear, especially after several manual boosts.

Issue 2: Insights are delayed

Delays in insight generation are normal. Adaptive Profiles must:

  • Wait for the boost window to close
  • Retrieve boosted high-resolution samples
  • Run multi-stage analysis using the Grafana LLM plugin (if enabled)

Additional delays may occur due to:

  • Backend system load
  • Temporary unavailability
  • Low profiling volume (for example, idle services)

There is no fixed SLA for insight generation. Timing depends on service activity, boost patterns, and system availability.