datadog
π―SkillInstallation
npx skills add g1joshi/agent-skills --skill datadogRun it in a terminal; it asks which AI coding tools to install it for (Claude Code, Codex, Cursor and more).
Uses skills.sh Β· Node.js 22.20+
After installing
Ask for related work and the agent uses it on its own, or call it yourself.
/datadogInstall by hand
Copy the whole skill folder from GitHub to the place below. Copying only SKILL.md can leave out files the skill uses.
Open the skill folder β- This project
.claude/skills/datadog/ - All projects
~/.claude/skills/datadog/
These paths are for Claude Code. For other tools, the command above puts it in the right place.
SKILL.md
View on GitHub βExpert Datadog monitoring assistance covering APM tracing, custom metrics, log collection, dashboards, and monitors. Use when instrumenting production infrastructure and observability.
Datadog
Datadog is a unified SaaS observability and security platform providing full-stack APM tracing, infrastructure monitoring, log aggregation, and real-time anomaly detection.
When to Use
- Enterprise Full-Stack Observability: Unified APM tracing, infrastructure metrics, log aggregation, and real-user monitoring (RUM).
- Distributed Microservice Tracing (APM): Visualizing request flow, latency bottlenecks, and database query timings across services.
- Automated Anomaly Detection & Alerting: Machine learning-powered alert monitors detecting traffic spikes, errors, and memory leaks.
- Cloud Cost & Security Posture Management: Tracking cloud resource misconfigurations, compliance, and container security.
Quick Start
1. Install Datadog Agent (Linux/macOS)
DD_API_KEY="<YOUR_DATADOG_API_KEY>" DD_SITE="datadoghq.com" \
bash -c "$(curl -L https://s3.amazonaws.com/dd-agent/scripts/install_script.sh)"
2. Auto-Instrument Node.js Application
npm install dd-trace --save
// At the absolute top of your entrypoint file (index.js)
import tracer from "dd-trace";
tracer.init({
service: "order-service",
env: "production",
logInjection: true,
runtimeMetrics: true,
});
Core Concepts
Auto-Instrumenting Applications with dd-trace
Adding distributed tracing to a Node.js microservice:
// Must be initialized before any other module imports!
import tracer from "dd-trace";
tracer.init({
env: "production",
service: "order-processing-service",
version: "2026.1.0",
logInjection: true, // Automatically inject trace IDs into logs
runtimeMetrics: true, // Collect GC and event loop latency metrics
});
export default tracer;
Custom Business Metrics & StatsD
Submitting custom business KPIs with high performance over UDP:
from datadog import initialize, statsd
initialize(statsd_host="127.0.0.1", statsd_port=8125)
# Increment transaction counter with tags
statsd.increment(
'ecommerce.checkout.success',
tags=['payment_method:apple_pay', 'region:us_east']
)
# Record latency histogram
statsd.histogram(
'ecommerce.checkout.duration_ms',
value=245.5,
tags=['tier:premium']
)
Correlating Logs with Distributed Traces
Formatting structured logs with active Datadog trace context:
{
"timestamp": "2026-03-27T10:15:30.123Z",
"level": "error",
"message": "Payment authorization timed out with gateway",
"dd": {
"trace_id": "847291038472918374",
"span_id": "1928374650192834",
"service": "order-processing-service",
"version": "2026.1.0",
"env": "production"
}
}
Common Patterns
Distributed APM Tracing with Custom Spans
Problem: Pinpointing exact latency bottlenecks within deeply nested asynchronous service methods.
Solution: Wrap business logic in custom Datadog tracer spans:
import tracer from "dd-trace";
tracer.init(); // Initialize tracer before importing other packages
async function processOrder(orderId) {
return tracer.trace(
"order.process",
{ resource: "checkout" },
async (span) => {
span.setTag("order.id", orderId);
const inventory = await checkInventory(orderId);
span.setTag("inventory.status", inventory.status);
const charge = await chargeCard(orderId);
return { inventory, charge };
},
);
}
Best Practices
Do:
- Enforce Unified Service Tagging across all systems:
env,service, andversiontags must match everywhere. - Enable
logInjection: truein APM tracers to link logs and traces together automatically in the UI. - Use DogStatsD over local UDP to submit high-frequency metrics without blocking application request threads.
- Configure log sampling and exclusion filters in
datadog.yamlto control ingestion costs.
Don't:
- Initialize APM tracers after importing HTTP or database modules; tracer must load first to patch libraries.
- Emit high-cardinality tags (e.g. user IDs, order UUIDs, timestamps) as custom metric tag dimensions; use logs/spans.
- Log sensitive PII or credentials; configure Datadog log scrub rules (
log_processing_rules).
Troubleshooting
| Error | Cause | Solution |
|---|---|---|
Datadog Agent: API Key invalid | DD_API_KEY contains typo or belongs to wrong Datadog site (e.g. EU vs US). | Verify API key and ensure DD_SITE is configured correctly (datadoghq.com vs datadoghq.eu). |
Traces not showing in APM dashboard | Application tracer unable to reach Datadog agent on port 8126. | Set DD_AGENT_HOST=127.0.0.1 and verify agent container port mapping. |
Too many metrics: Custom metric cardinality explosion | Using dynamic UUIDs or timestamps as custom metric tag values. | Move high-cardinality values from metric tags to log attributes. |
References
More from this repository10
A template skill providing a starter framework for creating new agent skills with structured sections for triggers, quick start examples, and capability descriptions.
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill