What are metrics and why are they needed?

Metrics are quantitative indicators used to measure and monitor the performance, status, and behavior of an application or system.

They are needed for:

• Performance assessment: Measuring response time, throughput, resource usage (CPU, memory, etc.).
• Problem detection: Identifying anomalies, failures, bottlenecks, and other potential issues.
• Status monitoring: Tracking service availability, error counts, component health.
• Decision making: Analyzing data for optimization, scaling, and development planning.
• SLO/SLA compliance: Checking adherence to established service levels.

Examples of key metrics:

• RPS (Requests Per Second) - Number of requests per second.
• Latency - Request execution time (e.g., p95, p99).
• Error Rate - Percentage of erroneous requests.
• CPU Usage - CPU utilization percentage.
• Memory Usage - Amount of used RAM.
• Disk I/O - Read/write operations on disk.
• Network Traffic - Volume of transmitted data.
• Queue Length - Length of the request processing queue.

Metrics are collected by monitoring tools (Prometheus, Grafana, Zabbix, etc.) and are used to create graphs, alerts, and dashboards.