Monitoring and Service Health

This guide describes the observability capabilities of Digital Retail Engine and how to verify service health.

Health Check Endpoint

Digital Retail Engine provides a health check endpoint that reports the status of all critical components:

GET /health

The response includes the status of:

Component	Description
Application	Overall service availability
Database	SAP HANA Cloud connectivity and response time
Cache	Redis connectivity and response time

Example Response

{
  "status": "UP",
  "version": "1.0.0",
  "uptime": 12345,
  "timestamp": "2026-03-27T10:30:00Z",
  "service": "digital-retail-engine",
  "checks": {
    "database": {
      "status": "UP",
      "responseTime": 12
    },
    "redis": {
      "status": "UP",
      "responseTime": 3
    },
    "webhookDelivery": {
      "status": "active"
    }
  }
}

The endpoint always returns HTTP 200 for the Cloud Foundry liveness probe. When the database or Redis is unreachable, the top-level status reports DEGRADED (not UP) while still returning 200; the affected check reports DOWN. The webhookDelivery check reports active only when Redis is configured and reachable (the webhook outbox worker depends on it), otherwise inactive; it is informational and never degrades overall status.

Service Level Objectives

The Digital Retail Engine service is operated with the following targets:

Metric	Target	Description
Availability	99.5%	Monthly uptime target
POS API Latency (P95)	< 500ms	95th percentile response time for promotion evaluation
POS API Latency (P99)	< 2s	99th percentile response time for promotion evaluation
Error Rate	< 1%	HTTP 5xx error rate

Available Metrics

Digital Retail Engine exposes Prometheus-compatible metrics at the /metrics endpoint. These metrics can be integrated into your existing monitoring infrastructure.

Application Metrics

Metric	Description
http_requests_total	Total HTTP requests processed (by method, route, status code)
http_request_duration_seconds	Request latency distribution
http_requests_in_flight	Currently active requests

Business Metrics

Metric	Description
evaluations_total	Total promotion evaluations (by tenant, status)
evaluate_duration_seconds	Duration of promotion evaluation (by tenant)
promotions_applied_total	Number of promotions applied to baskets
cache_hit_total	Cache hits for active promotions and evaluation results
cache_miss_total	Cache misses for active promotions and evaluation results

Integration Metrics (Public API + DRFOUT)

These metrics provide visibility into bulk imports, webhook delivery, and DRFOUT ingestion.

Metric	Description
dre_imports_total	Total imports processed (by tenant, type, status)
dre_imports_duration_seconds	Import job duration in seconds (by tenant, type)
dre_webhook_deliveries_total	Webhook deliveries (by tenant, status)
dre_webhook_dead_letters_total	Webhook dead-letter transitions (by tenant)
dre_drfout_skipped_total	DRFOUT items skipped (by tenant, reason)

Integrating with Your Monitoring Stack

The /metrics endpoint returns data in Prometheus text exposition format and can be scraped by any Prometheus-compatible monitoring tool (e.g., Grafana, Datadog, SAP Cloud ALM).

SAP Cloud ALM

For enterprise-grade monitoring with SAP Cloud ALM:

1. Enable Real User Monitoring and Health Monitoring capabilities in your SAP Cloud ALM tenant.
2. Configure the data collector to scrape the /metrics endpoint.
3. Set up Situation Automation rules for alerting on error rate and latency thresholds.

Troubleshooting

Service Appears Slow

1. Check the /health endpoint to verify database and cache connectivity.
2. Review the evaluate_duration_seconds metric — high values may indicate complex promotion configurations or a large number of active promotions.
3. Check the cache hit rate — a low hit rate increases database load. Cache is invalidated when promotions are modified; frequent updates may reduce cache effectiveness.
4. Contact DRE support if latency persists beyond the published SLOs.

Service Unavailable

1. Verify the /health endpoint returns a response.
2. Check the SAP BTP cockpit for any platform-level incidents.
3. Contact DRE support with the health check response and timestamp for investigation.