APM & Application Performance Analyst

Analyze application performance using APM tools like Datadog, New Relic, Dynatrace, and Elastic APM. Identify bottlenecks, tune instrumentation, and optimize service health.

Application Performance Monitoring tools generate enormous amounts of data — transaction traces, error rates, apdex scores, throughput metrics, dependency maps — but extracting actionable insights from that data requires both platform expertise and a systematic analytical approach. The APM and Application Performance Analyst helps engineering teams get maximum value from their APM investment, whether they are using Datadog, New Relic, Dynatrace, Elastic APM, or AppDynamics.

This assistant helps you navigate APM platforms to investigate specific performance problems: slow transactions, error spikes, degraded Apdex scores, database query bottlenecks, external dependency timeouts, memory leaks revealed by heap profiling, and CPU hotspots identified through continuous profiling data. When you describe a symptom, it guides you through the diagnostic workflow specific to your APM platform — which views to look at, which filters to apply, what the trace data is telling you, and how to move from a symptom in the APM dashboard to a root cause in the code or infrastructure.

Beyond incident investigation, the assistant helps you configure APM agents for optimal coverage and minimal overhead. It advises on custom instrumentation for business transactions, setting appropriate sampling rates, configuring alert thresholds for Apdex degradation and error rate spikes, and designing service maps that accurately reflect your architecture.

The assistant also helps teams compare APM platforms when evaluating a migration or a new deployment, understanding the trade-offs between agent-based and eBPF-based instrumentation, between full-stack versus application-only APM, and between commercial APM and open-source alternatives built on OpenTelemetry.

Ideal users include SREs and developers investigating production performance degradation, platform engineers evaluating or migrating APM tools, engineering managers who need to understand what their APM data is actually telling them, and teams onboarding a new APM platform and wanting to configure it correctly from the start.

🔒 Unlock the AI System Prompt

Sign in with Google to access expert-crafted prompts. New users get 10 free credits.

Sign in to unlock