AI Observability Summaries

BlameTrail can automatically generate AI-powered incident summaries by analyzing your connected metrics, logs, and deploy data. When an incident occurs, the AI correlates signals across providers to produce an actionable analysis.

What you get

Each AI summary includes:

• Summary -- A concise 2-4 sentence overview of what happened and when.
• Likely causes -- Ranked list of probable causes, each tied to specific evidence from your data, with a confidence level.
• Suspicious commits -- Commits flagged as potentially related, based on deploy timing and observed behavior changes.
• Observed anomalies -- Key metric and log anomalies detected around the incident window.
• Recommended actions -- Practical debugging and mitigation steps.
• Overall confidence -- How confident the AI is in its analysis (high, medium, or low), based on the strength of available signals.

When summaries are generated

Summaries are generated automatically when:

• An incident is opened (initial analysis)
• An incident is resolved (updated analysis with full timeline)
• New suspect scoring data arrives (deploy correlation)
• New deploy commit enrichment completes (richer context)

You can also regenerate a summary manually from the incident detail page, or run an ad-hoc analysis for any service and time range from the observability metrics page.

Where to find them

• Incident detail page -- The "AI Observability Summary" card appears in the Observability tab when a summary is available.
• Observability metrics page -- Use the "Analyze with AI" button to run ad-hoc analysis for a specific time range.

Plan limits

Disabling AI features

Organization owners can disable all AI features for their organization from the organization settings. When disabled, no AI summaries will be generated and ad-hoc analysis will be unavailable.