Skip to content

13.3 Metrics, Dashboards, Alerts

Sprint: Sprint 13 - Observability and Resilience
Feature ID: 13.3
Subtasks: 13.3.1, 13.3.2, 13.3.3

Objective

Deliver "Metrics, Dashboards, Alerts" within Sprint 13 - Observability and Resilience.

Dependencies

Aggregated from the subtasks below (prerequisite task IDs and enablers):

  • Sprints 04-12
  • 13.3.1
  • 13.3.2

Subtasks

13.3.1 Metrics exposure and Prometheus scraping

  • ID: 13.3.1
  • Title: Expose Micrometer metrics and scrape with Prometheus
  • Description: Ensure every service exposes /actuator/prometheus (HTTP latency, JVM, Kafka consumer lag, Resilience4j metrics) and Prometheus scrapes them (NFR-09).
  • Business Purpose: Quantitative health and performance signals (NFR-09).
  • Inputs: ADR-012, NFR-09.
  • Outputs: Metrics endpoints + Prometheus scrape config.
  • Acceptance Criteria:
  • Prometheus lists all services as up and collects HTTP, JVM, and Resilience4j metrics.
  • Dependencies: Sprints 04-12
  • Complexity: M

13.3.2 Grafana dashboards

  • ID: 13.3.2
  • Title: Provision Grafana dashboards for platform and domains
  • Description: Provision dashboards for API latency (p50/p95/p99 vs NFR-01), error rates, Kafka consumer lag, saga throughput, bill-run duration (NFR-02), and circuit-breaker state.
  • Business Purpose: Operoverview of platform health and NFR adherence (NFR-01, NFR-02, NFR-09).
  • Inputs: ADR-012, NFR-01, NFR-02.
  • Outputs: Provisioned Grafana dashboards.
  • Acceptance Criteria:
  • Dashboards render with live data; p95 latency and bill-run duration panels exist and populate.
  • Dependencies: 13.3.1
  • Complexity: M

13.3.3 Alert rules

  • ID: 13.3.3
  • Title: Define alerting rules for SLO breaches
  • Description: Prometheus/Grafana alert rules for p95 > 300ms sustained, error-rate spikes, consumer lag growth, and breaker-open state, aligned to the 99.5% uptime objective (NFR-01, NFR-04).
  • Business Purpose: Proactive detection of SLO violations (NFR-01, NFR-04).
  • Inputs: NFR-01, NFR-04.
  • Outputs: Alert rule definitions.
  • Acceptance Criteria:
  • A simulated latency/error breach fires the corresponding alert.
  • Dependencies: 13.3.2
  • Complexity: S

Acceptance Criteria

Each subtask above carries its own objective, testable acceptance criteria. This feature is complete when every subtask's acceptance criteria are met.

Deliverables

  • Metrics endpoints + Prometheus scrape config.
  • Provisioned Grafana dashboards.
  • Alert rule definitions.

Definition of Done

  • All subtasks (13.3.1, 13.3.2, 13.3.3) completed and their acceptance criteria verified.
  • All listed dependencies satisfied.
  • Build and tests green per the global conventions in ../README.md (Section 3).