Skip to main content

ADR-009: Logging and Monitoring Infrastructure

Status: Accepted

Context​

As the photobooth application grows in complexity with payment integrations, multiple printer support, and cloud storage, we need robust logging and monitoring to:

  • Debug production issues quickly
  • Track payment transaction health
  • Monitor printer connectivity and status
  • Observe system performance and capacity

Decision​

We will implement a multi-layer observability stack:

  1. Application Logging: Structured JSON logs with context (tenant_id, session_id, transaction_id)
  2. Error Tracking: Sentry for exception capture and grouping
  3. Metrics & Alerts: OpenTelemetry + Prometheus/Grafana Cloud (free tier)
  4. Uptime Monitoring: UptimeRobot or similar for endpoint monitoring

Architecture​

Logging Strategy​

Log Levels​

LevelUsage
ERRORPayment failures, printer errors, exceptions
WARNRetries, degraded performance, invalid inputs
INFOKey events (session start/end, payment success)
DEBUGDetailed flow for debugging

Required Context Fields​

interface LogContext {
  tenant_id?: string;
  session_id?: string;
  transaction_id?: string;
  printer_id?: string;
  user_id?: string;
  timestamp: string;
  environment: "development" | "production";
}

Monitoring Targets​

1. Payment Flow​

  • Payment creation success/failure rate
  • Webhook delivery success
  • Transaction processing time
  • QRIS scan to payment confirmation latency

2. Hardware/Printing​

  • Printer connection status
  • Print job success/failure rate
  • Average print time
  • Bluetooth disconnect events

3. System Health​

  • API response times
  • Error rates by endpoint
  • Cloud storage upload success
  • Session completion rate

Tools Selection​

CategoryToolTierCost
LoggingGrafana Cloud (Loki)Free$0
Error TrackingSentryTeam$0 (free tier)
MetricsGrafana Cloud (Prometheus)Free$0
UptimeUptimeRobotFree$0

Note: All selected tools have generous free tiers suitable for MVP.

Implementation Phases​

Phase 1: Foundation​

  • Integrate Sentry for frontend/backend error tracking
  • Add structured logging to payment flow
  • Create error boundary in React

Phase 2: Metrics​

  • Add OpenTelemetry instrumentation
  • Set up Grafana Cloud dashboard
  • Create payment success/failure charts

Phase 3: Alerts​

  • Configure error alerts (Slack/email)
  • Set up printer offline notifications
  • Payment failure webhook alerts

Consequences​

Positive​

  • Faster debugging in production
  • Data-driven optimization decisions
  • Proactive monitoring before users report issues
  • All tools have free tiers

Negative & Risks​

  • Additional DevOps complexity
  • Log storage costs at scale
  • Need to balance log verbosity vs. cost