RED and USE Methods
You'll learn two powerful frameworks for deciding what metrics to monitor in production systems.
The Problem They Solve
After learning the Four Golden Signals, you might wonder: "Which metrics apply to what parts of my system?" The RED and USE methods answer this by providing focused frameworks for different monitoring scenarios.
The RED Method (Request-Centric)
RED focuses on services and endpoints—anything that handles requests:
- Rate: How many requests per second?
- Errors: How many of those requests failed?
- Duration: How long do requests take?
Think of RED as monitoring from your user's perspective. If you're tracking an API endpoint, RED tells you if users are experiencing slowness (duration), failures (errors), or if traffic is spiking (rate).
Example: Your /checkout endpoint might show: 150 req/s (Rate), 2% error rate (Errors), 450ms p99 latency (Duration).
The USE Method (Resource-Centric)
USE focuses on physical resources—hardware and infrastructure:
- Utilization: How busy is this resource? (percentage)
- Saturation: How much work is queued/waiting?
- Errors: Are there resource-level errors?
Think of USE as monitoring your system's internal health. It answers "Is my CPU/disk/memory causing problems?"
Example: Your database server might show: 75% CPU utilization, 20 queries waiting in queue (saturation), 3 disk read errors.
When to Use Which
- RED: Apply to services, APIs, microservices, endpoints
- USE: Apply to CPUs, disks, network interfaces, memory, database connections
Key Takeaway: RED monitors request flows from the user's viewpoint, while USE monitors resource health from the infrastructure's viewpoint—together they provide complete observability coverage.