The USE Methodology
USE stands for Utilization, Saturation, and Errors. Developed by Brendan Gregg, it’s a low-level diagnostic tool for infrastructure and system resources.
-
Utilization – How much of a resource is used? (e.g., 70% CPU)
-
Saturation – Is the system over capacity? (e.g., CPU run queue)
-
Errors – Are there hardware or system faults? (e.g., disk errors)
Best for: Bottleneck analysis on servers, network, disk, memory
The RED Methodology
RED focuses on service-level performance and is tailored for microservices and HTTP-based systems.
-
Rate – Requests per second
-
Errors – Failed requests
-
Duration – Request latency
Best for: APIs, user-facing services, and SLAs
The Four Golden Signals
Popularized by Google’s SRE book, this approach extends RED by adding traffic and resource saturation:
-
Latency
-
Traffic
-
Errors
-
Saturation
Best for: General-purpose service monitoring, especially in SRE and DevOps
Service Level Indicator, Service Level Objective and Service Level Agreement
These terms help teams align technical monitoring with business reliability goals:
-
SLI (Indicator) – A measurable metric (e.g., success rate)
-
SLO (Objective) – A reliability target (e.g., 99.9% uptime)
-
SLA (Agreement) – A formal contract with customers
Best for: Defining acceptable downtime and aligning reliability with business impact
The Three Pillars of Observability
Rather than focusing on what to measure, this model focuses on data types:
-
Logs – Structured/unstructured records
-
Metrics – Time-series measurements
-
Traces – End-to-end request flows
Best for: Designing your observability stack and tooling (e.g., OpenTelemetry)
When to Use What?
| Methodology | Focus | Layer | Best For |
|---|---|---|---|
USE |
System resources |
Infrastructure |
Debugging bottlenecks |
RED |
Request performance |
Applications/APIs |
User experience & alerting |
Four Golden Signals |
Overall health |
Services + Infra |
SRE monitoring |
SLI/SLO/SLA |
Business alignment |
Application & Org |
Reliability targets & reporting |
Three Pillars |
Observability data |
All layers |
Tooling and platform design |