SLA / SLO / SLI

Service Level Management Metrics

1. Overview of SLI/SLO/SLA, the Core of Quantitative Service Level Management

    flowchart LR
    A["Ambiguous service quality"] -- "Shift to quantitative, metric-based agreements" --> B["SLI/SLO/SLA system"]

Definition: An integrated management system that measures a service’s performance and reliability with objective indicators (SLI), sets target values to be achieved (SLO), and defines the commitment made when those targets are missed (SLA).

Characteristics: (Aligned expectations) The SLA-SLO-SLI hierarchy clearly aligns expectations between the service provider and its users. (Error budget) Quantifies the acceptable range of failure with an Error Budget, managing the balance between development velocity and stability with data. (Measurement-based) Objectively measures and manages service quality through quantitative indicators (SLIs) such as availability, latency, and throughput.

2. Components and Mechanism of Service Level Management

a. The Indicator-to-Agreement Model

    flowchart TD
    SLI["SLI (Service Level Indicator)<br/>What are we measuring?<br/>(the actual measured metric)"]
    SLO["SLO (Service Level Objective)<br/>What is the target value?<br/>(the internal team goal)"]
    SLA["SLA (Service Level Agreement)<br/>What is the compensation for missing it?<br/>(the contract with the customer)"]

    SLI -->|"Quantify"| SLO
    SLO -->|"Formalize into a contract"| SLA

    style SLI fill:#E3F2FD,stroke:#1976D2
    style SLO fill:#FFF3E0,stroke:#F57C00,stroke-width:2px
    style SLA fill:#FFEBEE,stroke:#D32F2F

Term	Definition	Example
SLI	A quantitative measure of service level	Latency, availability, throughput
SLO	An agreed target value for the service level	99.9%+ availability, response time under 200ms
SLA	The financial/legal liability incurred when an SLO is missed	Service fee refund, credit issuance, right to terminate the contract

b. The Valet Model and Golden Signals for Deriving SLIs

    flowchart TD
  Root["Service Indicator Models"] --> Valet["Valet Model"]
  Valet --> Volume["Volume - traffic volume"]
  Valet --> Availability["Availability"]
  Valet --> Latency["Latency"]
  Valet --> Error["Error rate"]
  Valet --> Ticket["Ticket count"]
  Root --> SRE["Golden Signals - SRE"]
  SRE --> LatencySRE["Latency"]
  SRE --> TrafficSRE["Traffic"]
  SRE --> ErrorsSRE["Errors"]
  SRE --> Saturation["Saturation"]

Metric Type	Key Measurement Items	Notes
Availability	Ratio of successful requests among all requests	Confirms whether the service is up
Latency	Time taken to process a service request	Directly affects user experience (UX)
Errors	Frequency of explicit/implicit errors	Confirms system integrity
Saturation	Resource usage relative to its limit	Basis for judging scalability

3. Expected Benefits and Practical Strategy of SLA / SLO / SLI

Category	Expected Benefits	Application and Practical Use
Decision-making standard	Use of the Error Budget	Decide whether to ship new features based on SLO attainment
Transparency	Makes service reliability visible	Share status via real-time dashboards (e.g., Grafana)
Clear R&R	Defines responsibility between provider and user	Sets response priority and recovery time objective (RTO) during an incident
Continuous improvement	Data-driven service enhancement	Periodic SLO review to optimize over- or under-provisioned infrastructure

Six Sigma Cost of Quality (CoQ)