Course contentsShow
← Cloud and Infrastructure
Lesson 632 of 2,98912. Monitoring FoundationsFree lesson

Application Metrics

Application-level metrics: request rates, response times, error rates, and business KPIs.

Application Metrics

You'll learn what application-level metrics reveal about your software's health and user experience.

What Are Application Metrics?

While infrastructure metrics tell you how your servers are doing (CPU, memory, disk), application metrics tell you how your software is performing from a user and business perspective. These are measurements captured from within your running application code.

Core Application Metrics

Request Rates

The volume of incoming requests your application handles over time. Are you processing 100 requests per second or 10,000? This tells you about user activity and load patterns.

Analogy: Like counting how many customers walk through your store's entrance each hour.

Response Times

How long it takes your application to fulfill a request. Often measured as averages (mean) or percentiles (p50, p95, p99). A p95 of 200ms means 95% of requests complete within 200 milliseconds.

Analogy: The time between a customer placing an order and receiving their meal.

Error Rates

The percentage of requests that fail. This includes HTTP 5xx errors, exceptions, timeouts, or failed database queries. A sudden spike usually means something broke.

Analogy: How many orders come out wrong in your kitchen.

Business KPIs

Metrics specific to your application's purpose: successful logins, items added to cart, payments processed, files uploaded. These connect technical performance to business outcomes.

Why They Matter

Application metrics answer questions infrastructure metrics can't: "Are users experiencing slow checkouts?" "Did that deployment increase error rates?" "How many customers completed their signup?"

Key Takeaway: Application metrics measure your software's behavior from the user's perspective—request volumes, speed, errors, and business outcomes—complementing infrastructure-level monitoring.