Monitoring

Overview

Even when a service appears to be running smoothly, you never know when a problem will occur.

• You need a way to proactively check its health and quickly detect issues when they arise

• Key performance indicators
  • API performance metrics
    • It’s common practice to set a performance target for your API
  • SLA (Service Level Agreement)
    • A formal contract with your users
    • Defines the level of service you promise to deliver, including crucial metrics like uptime (e.g., 99.9% availability) and guaranteed response times
• The primary goals of monitoring are:
  • Early Failure Detection: Find and fix problems before they impact users.
  • Identify Performance Bottlenecks: Pinpoint exactly what part of the system is slow.
  • Prevent Resource Waste: Ensure you are using your infrastructure efficiently.
  • Analyze Business Trends: Understand how usage patterns change over time.
  • basically health checks and logging & analysis

Tools

A combination of tools is typically used to build a robust monitoring system.

Tool	Role
Prometheus	Collects and stores time-series metrics from your application.
Grafana	Creates visualization dashboards from data sources like Prometheus and triggers alerts based on metrics.
Spring Boot Actuator	Exposes critical JVM and application health information (like memory usage and HTTP request data) via HTTP endpoints.
Micrometer	An application metrics facade that allows your Spring application to send metrics to various monitoring systems (like Prometheus) without being locked into a specific one.

• Real world apps use Prometheus and Grafana the most
  • Spring Boot Actuator are only used in Spring
• If you don’t have a dashboard, you can just search in the terminal using grep
  • grep - searches files for lines containing a specific pattern
  • use it to quickly find errors, trace a specific user’s activity, or monitor live logs in real-time
  • Finding all error messages in a log file: grep "ERROR" application.log
• Spring Boot Actuator