SLI
A quantitative metric framework for measuring service health, encompassing dimensions such as successful request counts, throughput, and failure rates. SLIs are a core tool in Site Reliability Engineering (SRE) for defining and monitoring service quality, and serve as the basis for establishing Service Level Objectives.
Timeline (last 90 days)
According to the Google SRE system, an SLI (service level indicator) is a precise quantitative measure of some aspect of a service.
An SLI is the actual measured value, an SLO is the internal target the team sets for the measured value, and the error budget is the room for failure allowed by 100% minus the SLO; the three form a closed loop.
Five steps for setting reliability goals for AI features: choose only one quality SLI per feature, version the eval set, set the SLO from a week of baseline data, tag AI-written PRs, and define explicit failure blocking rules.
Approximately five years ago, Azure established the principle that a service is only considered healthy when customers perceive it as healthy, and defined a standardized Service Level Indicator (SLI) architecture that requires all services to report to Azure Monitor according to a unified schema.
There are three industry implementation approaches for automatically inferring which metrics to monitor: static analysis based on code diffs, establishing baselines from historical deployment data for comparison, and engineers declaring change intent using a DSL or configuration.
All Facts (5)
According to the Google SRE system, an SLI (service level indicator) is a precise quantitative measure of some aspect of a service.
50%UnverifiedAn SLI is the actual measured value, an SLO is the internal target the team sets for the measured value, and the error budget is the room for failure allowed by 100% minus the SLO; the three form a closed loop.
50%UnverifiedFive steps for setting reliability goals for AI features: choose only one quality SLI per feature, version the eval set, set the SLO from a week of baseline data, tag AI-written PRs, and define explicit failure blocking rules.
50%UnverifiedApproximately five years ago, Azure established the principle that a service is only considered healthy when customers perceive it as healthy, and defined a standardized Service Level Indicator (SLI) architecture that requires all services to report to Azure Monitor according to a unified schema.
50%UnverifiedThere are three industry implementation approaches for automatically inferring which metrics to monitor: static analysis based on code diffs, establishing baselines from historical deployment data for comparison, and engineers declaring change intent using a DSL or configuration.
50%