● generator operational EN FR EN

ninesarewelcom

Service objectives, written like a manual.

About Services

Articles ▾

User Guide

SLI & SLO Error Budget Burn Rate

Alert Noise Prometheus Cardinality MTTR & MTTF Realistic SLO Dependency Risk

Simulators ▾

All simulators

SLO Simulator Realistic SLO Alert Noise Reducer Cardinality Budget MTTR/MTTF Optimizer Dependency Risk Scorer

Changelog Plans Register Login

About Services User Guide All simulators Changelog Plans

Step 1 — Service & SLI Definition

Define your service and the kind of SLI you want to track.

1 · Service & SLI

2 · Metric Wiring

3 · SLO & Window

4 · Review & Generate

Quick-start presets

optional

REST API availability HTTP 2xx/3xx ratio for web APIs and microservices gRPC error rate Error code ratio for gRPC services Blackbox probe External uptime monitoring via Blackbox Exporter Kafka consumer Throughput SLO for message queue consumers Data pipeline freshness Freshness gauge for ETL jobs and data pipelines

Service

Service name *

Identifies the service this SLO applies to. Used in metric labels.

SLO name *

Used as prefix for recording rule names. Use snake_case, no spaces.

SLI Kind

III

SLI Kind

Web services & APIs

Availability Ratio of successful requests to total requests. Latency Ratio of requests served within a latency threshold. Error Rate Ratio of error responses to total requests. Throughput Volume of requests processed per time unit. Probe Blackbox probe success ratio (e.g. Blackbox Exporter).

Quality & compliance

Quality Fraction of requests served at full quality (no degraded fallback). Correctness Fraction of outputs that are correct (pipelines, batch, calculations).

Data & pipelines

Coverage Fraction of expected records actually processed (ETL, batch ingestion). Durability Fraction of data confirmed intact over time (storage, object stores). Freshness Fraction of time data was fresh (binary gauge: 1=fresh, 0=stale).

SLI Type

Request-based Measured per individual request. Most common. Uses rate(). Window-based Measured over fixed time windows (e.g. 1m probes). Uses avg_over_time().

Ready-to-deploy artifacts: Prometheus recording rules, multi-burn-rate alerts, Grafana dashboards, runbooks.

About Services Changelog Manual Plans

ninesarewelcom an arewel workshop · designed in Belgium