Olly mini
Olly mini is Coralogix's embedded AI helper. It powers focused, single-purpose AI actions inside specific product screens — analyzing a flame graph, turning a plain-English prompt into a query, summarizing a case, or explaining what happened in a log entry or error — so you can act on the data in front of you without switching context.
Olly mini complements Olly, the conversational AI observability agent. Where Olly investigates open-ended questions through chat across all your telemetry, Olly mini stays in-page and returns a single answer scoped to what you're looking at.
You can recognize an Olly mini result by its panel design: a collapsible box with the answer inside, a thumbs up / thumbs down control, and a footer noting that the content was generated by Olly mini.
Where Olly mini appears
| Feature | Where to find it | What it does |
|---|---|---|
| Query assistant | Explore logs or Explore spans, DataPrime toggle → Query Assistant | Converts a plain-English prompt into a DataPrime query. |
| Explain log | Any page that lists logs → hover a log entry → Explain log | Explains what a single log entry reports — its severity, impact, and likely cause. |
| Analyze flame graph | APM → Service Catalog → Profiles → flame graph → Analyze | Highlights hotspots, root causes, and optimization suggestions in a CPU or wall-clock flame graph. |
| Analyze compare flame graph | Continuous Profiling Compare mode → Analyze | Explains what changed between profile A and profile B — regressions, improvements, and likely causes. |
| Analyze errors | RUM → Error Tracking → error template drilldown → Analyze | Analyzes a single error occurrence — stack trace, error message, and runtime context — and returns a probable cause, debugging steps, and next actions. |
| AI summary in Cases | Any case → Triage tab | Generates a plain-language summary of the case from its evidence. Select Regenerate to refresh it. |
| Claude session analysis in AI Center | AI Center → Code agents → Claude dashboard | Summarizes a Claude Code session — activity, cost, and context. |
How Olly mini differs from Olly
- Olly is a conversational observability agent for open-ended investigation. Open a chat and ask questions across your logs, metrics, traces, and alerts.
- Olly mini is embedded in specific product surfaces to do a single, focused job on the data already in front of you. There is no conversation — you invoke it, read the answer, and continue where you were.
You do not enable Olly mini separately. When you enable Olly for your team, the Olly mini surfaces across the product become available. Some individual features still expose their own on/off toggle — for example, Query Assistant and Explain log each have a toggle under Settings → Account Preferences → AI-Powered Capabilities.
Billing
Olly mini usage is metered under the AI Units billing model, alongside Olly chat. Each invocation is charged per token at the model's rate. Free plans include 10 AI Units per user per month, plus a team-level cap of 100 AI Units per month across all members.
To track your Olly mini consumption, open the AI Units tab on the Data Usage page and review the Olly breakdown.
Feedback
Every Olly mini panel includes a thumbs up and thumbs down control. Use it to signal whether the answer was useful — feedback is aggregated to improve future results across the product.
Next steps
- Enable Olly for your team to activate Olly mini across the product.
- Review pricing in AI Units pricing.