# Observability default configuration.
# See docs/08-operations-observability/observability-strategy.md and metrics-slos-and-alerts.md.

tracing:
  samplingRatioDefault: 0.1
  samplingRatioSensitiveActions: 1.0     # approvals, employee conversion always fully sampled
  exporterEndpointEnvVar: OTEL_EXPORTER_OTLP_ENDPOINT

metrics:
  backend: prometheus                    # configurable: prometheus | azure_monitor | datadog
  scrapeIntervalSeconds: 30

logging:
  level: Information
  redactionEnabled: true                 # must remain true — see logging-and-redaction-standard.md
  structuredFormat: json

slos:
  apiAvailabilityPercent: 99.9
  cvParseLatencyP95Seconds: 60
  matchingRunLatencyP95Seconds: 120
  ragRetrievalLatencyP95Seconds: 3
  offerDeliveryLatencyP95Hours: 1
  employeeConversionLatencyP95Minutes: 15

alerting:
  apiErrorRateThresholdPercent: 5
  slaBreachEscalationEnabled: true
  budgetAlertThresholdsPercent: [80, 100]

costAllocationTags:
  - tenantId
  - modelVersionId
  - workflowStage
  - featureFlagName