Last updated:
Get Live Insights
Live Insights surfaces active problems in your environment before you need to look for them. The Agent0 landing page at app.dash0.com/agent0 surfaces what is actively going wrong in your environment, refreshed every 60 seconds, grouped per service.
Each Live Insight entry includes the affected service, the signal that triggered it, a severity rating (critical / warning / info), and a suggested investigation prompt.
Signal Categories
Live Insights monitors and surfaces the following types of issues:
- Failing checks — Services with active failing checks, ordered by severity.
- Elevated error rates — Services whose error rate has spiked relative to their baseline.
- Elevated latency — Services whose p95 latency has spiked relative to their baseline.
- Pod saturation — Services with pods under memory or CPU pressure.
- Recent deployments — Services deployed in the last 30 minutes, with a comparison prompt to check for regressions.
- Slow database calls — Services with database operations that have gotten significantly slower.
Severity Bands
For error-rate and p95-latency insights, severity compares the current value with the corresponding calculation window 24 hours earlier:
- Critical: Current value is at least 5× the historical value.
- Warning: Current value is at least 2× but less than 5× the historical value.
- Info: Current value is below 2× the historical value but meets the criteria for inclusion.
This comparison uses an equivalent window 24 hours earlier, not an aggregate of the previous day. If the historical value is missing or unusable, absolute thresholds determine severity instead. A service can appear without historical data. When a usable historical ratio exists, absolute thresholds do not override its severity.
Each of these two probes includes up to five services with at least ten requests in the calculation window. Error-rate candidates must have an error rate of at least 1%; latency candidates must have a p95 latency of at least one second.
Start an investigation
Select an insight to fill the Chat composer with a prompt and context scoped to the issue. Review the draft, then send it to start the investigation. Selecting an insight alone does not run Agent0.
How Live Insights Works
Live Insights continuously monitors your telemetry data and applies statistical analysis to detect anomalies. The system:
- Evaluates every 60 seconds — Fresh insights based on the latest telemetry.
- Groups by service — All issues for a service appear together for easier triage.
- Prioritizes by severity — Critical issues appear first, followed by warnings and info-level insights.
- Provides context — Each insight includes the specific metric, threshold, and current value.
If a service doesn't appear in Live Insights, it may have no recent telemetry, fall below a probe's inclusion thresholds, or fall outside its result limit. See Live Insights Not Showing Services for troubleshooting.
Further Reading
- About Chat — Overview of Agent0 Chat features
- Investigation & Analysis — Perform multi-signal root cause analysis

