Live Performance

Core Charts

Performance charts use real Runtime metric buckets and show axes, tooltips, and time-range controls.

Key metrics include:

• Tick performance p50, p95, and p99.
• RTT and jitter.
• Acknowledgement lag.
• Throughput in and out.
• Replication efficiency.
• Backpressure events.
• Reject reason breakdown.

Regressions

Live compares current performance to a trailing per-deployment baseline. A sustained 25 percent degradation is a warning. A sustained 50 percent degradation is critical.

Regression rows show:

• Severity.
• Current value.
• Baseline value.
• Change percentage.
• Sustained bucket count.
• Recommended action.
• Notification status.

Critical regressions appear in Live immediately. Email delivery for critical regressions is pending the notification hookup listed in the shipped audit.

Reading A Spike

Use the chart marker and the activity feed together. A one-bucket spike may be noise. Sustained markers usually mean a room, deployment, script, or replication path needs investigation.