spectral-edge-rules.yml

# Example Prometheus alerting rules for spectral-edge (copy into your Prometheus or use as a starting point).
# Adjust thresholds to your SLOs. Requires metrics from /metrics (or -metrics-listen).

groups:
  - name: spectral-edge.rules
    rules:
      - alert: SpectralEdgeUpstreamErrors
        expr: |
          sum(rate(spectral_edge_upstream_errors_total[5m])) > 0.05
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "spectral-edge upstream round-trip failures"
          description: "Upstream transport errors exceed 0.05/s (5m rate). Check origin health and network."

      - alert: SpectralEdgeHighUpstreamLatency
        expr: |
          histogram_quantile(0.95,
            sum(rate(spectral_edge_upstream_roundtrip_duration_seconds_bucket[5m])) by (le)
          ) > 10
        for: 10m
        labels:
          severity: warning
        annotations:
          summary: "spectral-edge p95 upstream RTT high"
          description: "p95 upstream round-trip time is above 10s. Tune timeouts or upstream capacity."

      - alert: SpectralEdgePolicyAlertsSpike
        expr: |
          sum(rate(spectral_edge_policy_alerts_total[5m])) > 10
        for: 5m
        labels:
          severity: info
        annotations:
          summary: "spectral-edge policy alert rate elevated"
          description: "Policy matches logged at >10/s. May be expected during incidents—confirm with app owners."

      - alert: SpectralEdgeBodyTruncated
        expr: |
          sum(rate(spectral_edge_body_truncated_total[5m])) > 0.1
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "spectral-edge rejecting large bodies (413)"
          description: "Requests exceed -max-body; clients may need smaller payloads or raise -max-body with care."