Slo Sli Design

Skill: SLO/SLI Design

> Expertise: SLI selection, SLO target setting, error budget calculation, burn rate alerting, Sloth/pyrra integration.

When to load

When defining SLOs for a new service, setting up error budget tracking, or reviewing existing SLOs after an incident.

SLI Selection Framework

Step 1: What does the user care about?
  → "The checkout completes successfully and quickly"

Step 2: What CAN we measure?
  → HTTP 2xx responses, p99 latency

Step 3: Define the SLI formula
  → Availability SLI: good_requests / total_requests
     where good = status  14.4× rate)"
          runbook_url: "https://runbooks.internal/slo-fast-burn"

      # Slow burn: 3× rate (burns 10% of budget in 6h)
      - alert: SLOSlowBurn
        expr: |
          (
            slo:http_availability:ratio_rate6h{service="checkout-service"}  3× rate)"

Sloth (SLO as Code)

# slo/checkout-service.yaml — Sloth generates all recording rules + alerts
version: "prometheus/v1"
service: checkout-service
labels:
  team: backend
slos:
  - name: requests-availability
    objective: 99.5
    description: "99.5% of checkout requests succeed with latency < 500ms"
    sli:
      events:
        error_query: |
          sum(rate(http_requests_total{service="checkout-service", status=~"5.."}[{{.window}}]))
        total_query: |
          sum(rate(http_requests_total{service="checkout-service"}[{{.window}}]))
    alerting:
      name: CheckoutServiceSLO
      page_alert:
        labels: { severity: critical }
      ticket_alert:
        labels: { severity: warning }

# Generate Prometheus rules from Sloth definition
sloth generate -i slo/checkout-service.yaml -o prometheus-rules/slo-checkout.yaml

Error Budget Dashboard (Grafana)

Key panels:

1. SLI over 28 days — current ratio vs SLO target line
2. Error budget remaining — percentage + time remaining (burn at current rate)
3. Burn rate — 1h, 6h, 1d, 7d windows
4. Events causing budget consumption — top error causes by count

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: sawrus
• Source: sawrus/agent-guides
• License: MIT
• Homepage: https://sawrus.github.io/agent-guides

Install and usage instructions live in the source repository linked above.