logging-and-observability-principles

name	logging-and-observability-principles
description	Structured logging and observability patterns: log levels, structured fields, correlation IDs, language-specific logger setup.
user-invocable	false

Logging and Observability Principles

⚠️ All operations MUST be logged per Logging and Observability Mandate GEMINI.md § Logging and Observability Mandate.

Log Levels

Level	When	Examples
TRACE	Extremely detailed diagnostics	Fn entry/exit, var states (dev only)
DEBUG	Detailed flow for debugging	Query exec, cache hits, state transitions
INFO	General informational	Request started, task created, login
WARN	Potentially harmful	Deprecated API, fallback, retry
ERROR	Error, app continues	Request failed, API timeout, validation
FATAL	Severe, causes shutdown	DB unreachable, critical config missing

Rules

1. Every operation logs start + success/error:

log.Info("creating task",
"correlationId", correlationID,
"userId", userID,
"title", task.Title,
)

log.Info("task created successfully",
"correlationId", correlationID,
"taskId", task.ID,
"duration", time.Since(start),
)

log.Error("failed to create task",
"correlationId", correlationID,
"error", err,
"userId", userID,
)

2. Context fields: correlationId (UUID), userId, duration (ms), error.

3. Structured only:

// ✅ Structured
log.Info("user login", "userId", userID, "ip", clientIP)

// ❌ String formatting
log.Info(fmt.Sprintf("User %s logged in from %s", userID, clientIP))

4. Never log: passwords, API keys, tokens, credit cards, PII (sanitize if needed), full req/res bodies.

5. Never log in hot paths: tight loops, per-item batch processing, latency-critical sync paths. Use logger middleware redaction (pino-redact, zap masking).

Language Implementations

Go (slog):

import "log/slog"

logger := slog.New(slog.NewJSONHandler(os.Stdout, &slog.HandlerOptions{
  Level: slog.LevelInfo,
}))

logger.Info("operation started",
  "correlationId", correlationID,
  "userId", userID,
)

TypeScript (pino):

import pino from 'pino';

const logger = pino({
  level: process.env.LOG_LEVEL || 'info',
});

logger.info({
  correlationId,
  userId,
  duration: Date.now() - startTime,
}, 'task created successfully');

Python (structlog):

import structlog

logger = structlog.get_logger()

logger.info("task_created",
correlation_id=correlation_id,
user_id=user_id,
task_id=task.id,
)

Log Patterns

API: log request received (method, path, correlationId) + request completed (status, duration).

DB: DEBUG level — query start (query text) + complete (rows, duration) + error (query, error).

External API: INFO start (service, endpoint). WARN retry (attempt, error). WARN circuit breaker (failureCount).

Background Jobs: INFO start (jobId, type) + progress (periodic, not per-item) + complete (duration, items).

Errors: ERROR recoverable (sanitized input). FATAL critical dependency (action: shutting down).

Environment Config

Env	Level	Format	Destination
Development	DEBUG	Pretty (colored)	Console
Staging	INFO	JSON	Stdout → cloud logs
Production	INFO	JSON	Stdout → cloud logs

func configureLogger() *slog.Logger {
var handler slog.Handler

    level := slog.LevelInfo
    if os.Getenv("ENV") == "development" {
        level = slog.LevelDebug
        handler = slog.NewTextHandler(os.Stdout, &slog.HandlerOptions{
            Level: level,
        })
    } else {
        handler = slog.NewJSONHandler(os.Stdout, &slog.HandlerOptions{
            Level: level,
        })
    }
    
    return slog.New(handler)
    }

Testing Logs

Capture and assert on log output:

func TestUserLogin(t *testing.T) {
var buf bytes.Buffer
logger := slog.New(slog.NewJSONHandler(&buf, nil))

    service := NewUserService(logger, mockStore)
    err := service.Login(ctx, email, password)
    
    require.NoError(t, err)
    logs := buf.String()
    assert.Contains(t, logs, "user login successful")
    assert.Contains(t, logs, email)
    }

Monitoring

Correlation IDs: generate at ingress, propagate through all services, include in all logs/errors/traces. UUID v4.

Aggregation: centralized (CloudWatch/GCP/Datadog). Index: correlationId, userId, level, timestamp. Alert on ERROR/FATAL. Dashboard: request rates, error rates, latency.

Checklist

• All public ops log INFO on start
• All ops log INFO/ERROR on complete/failure
• All logs include correlationId
• No sensitive data
• Structured (key-value pairs)
• Appropriate level
• Error logs include error details
• Perf-critical paths use DEBUG

Related

• Logging and Observability Mandate GEMINI.md § Logging and Observability Mandate
• Monitoring and Alerting Principles @.gemini/skills/monitoring-and-alerting-principles/SKILL.md
• Error Handling Principles GEMINI.md § Error Handling Principles
• Security Mandate GEMINI.md § Security Mandate
• Security Principles GEMINI.md § Security Principles
• API Design Principles @.gemini/skills/api-design-principles/SKILL.md