This Claude skill provides comprehensive database observability by instrumenting SQL queries, monitoring connection pool health, and identifying critical performance bottlenecks. It enables developers and SREs to proactively detect N+1 query patterns, track connection wait times, and optimize slow database operations through standardized tracing and metrics collection.
Use Cases
Detecting N+1 query patterns in application code to reduce unnecessary database round-trips and improve overall response times.
Monitoring connection pool metrics such as active, idle, and wait counts to prevent service exhaustion and optimize resource allocation during peak traffic.
Identifying slow-performing SQL statements by analyzing query duration histograms and implementing necessary indexing or query optimizations.
Standardizing database instrumentation across microservices using query wrappers to ensure consistent observability and faster incident response.
name
database-observability
description
"Instrument database queries, connection pools, and detect N+1 queries"
priority
2
Database Observability
Database is often the bottleneck. Track queries, pools, and patterns.
Query Span Attributes
Attribute
Example
Required
db.system
postgresql, mysql
Yes
db.operation
SELECT, INSERT
Yes
db.name
orders_db
Yes
db.sql.table
users
Recommended
db.statement
SELECT * FROM users WHERE id = ?
Parameterized only!
Connection Pool Metrics (USE)
Metric
Type
Description
db.connections.active
Gauge
In use
db.connections.idle
Gauge
Available
db.connections.max
Gauge
Pool limit
db.connections.wait_count
Counter
Had to wait
db.connections.wait_duration
Histogram
Wait time
Issues to Detect
Issue
Detection
Fix
N+1 queries
>10 identical queries per request
Use eager loading
Slow queries
Duration > p95 threshold
Add indexes, optimize
Pool exhaustion
wait_count increasing
Increase pool, fix leaks
Query Wrapper Pattern
Before: Start span (db.system, db.operation, db.sql.table), start timer
After: Record duration, set db.rows_affected, record errors, end span
Pool Monitor Pattern
Every 10s: Record active, idle, wait_count, wait_duration from pool stats
Anti-Patterns
Full SQL with values → PII risk, use parameterized only
