SLO Document Template

This is the canonical per-service template referenced by the SLO Registry — every row in the registry links back to a filled-in copy of this document.

[Service Name] — SLO Document

• Service: [Full service name]
• Owner: [Team name]
• Service Tier: Tier 1 | Tier 2 | Tier 3
• Status: active | proposed | deprecated
• Created: YYYY-MM-DD
• Last Reviewed: YYYY-MM-DD
• Review Cadence: 90 days (or after any SLO miss event)
• Error Budget Dashboard: [FILL: Grafana dashboard URL]

---

1. Service Overview

2 sentences: what does this service do and who depends on it?

Example: order-fulfillment-api processes checkout and payment events for the e-commerce platform. It is a synchronous dependency for the checkout service and an async dependency for the inventory service.

---

2. SLIs (Service Level Indicators)

Define the measurable signals that represent the service’s reliability. Prefer request-based SLIs over resource-based — see the KPI → SLI → SLO → SLA framework for how these terms relate.

| SLI Name | PromQL / Query | Good Event Definition | Notes | | ------------------ | --------------------------------------------------------------------------------------- | ------------------------------------------------- | -------------------------------- | ------------------------------- | | Availability | sum(rate(http_requests_total{status=~"2.. | 3.."}[5m])) / sum(rate(http_requests_total[5m])) | HTTP 2xx + 3xx responses | Excludes health check endpoints | | Latency P99 | histogram_quantile(0.99, sum(rate(http_request_duration_seconds_bucket[5m])) by (le)) | < [N]ms | Measured at the service boundary | | [FILL: custom SLI] | [FILL: query] | [FILL: good event definition] | [FILL: notes] |

Excluded traffic:

• [FILL: health check paths, synthetic monitors, internal admin endpoints]

---

3. SLOs (Service Level Objectives)

SLI	Target	Measurement Window	Error Budget
Availability	[FILL: e.g. 99.9%]	30-day rolling	[FILL: e.g. 43.8 min/month]
Latency P99	[FILL: e.g. 99%] requests < [N]ms	30-day rolling	[FILL: e.g. ~7.2 hr/month of violations]

Rationale for targets:

• [FILL: why 99.9% and not 99.95%? Reference user impact data or dependency SLAs.]

---

4. Alerting Policy

Multi-window, multi-burn-rate alerts. Short windows catch fast burns; long windows catch slow burns.

Alert Name	Burn Rate	Window	Severity	Response
[ServiceName]AvailabilityCritical	14.4×	1h	SEV1	Page immediately
[ServiceName]AvailabilityHigh	6×	6h	SEV2	Page within 30 min
[ServiceName]AvailabilityWarning	1×	3d	SEV3	Ticket (no page)
[ServiceName]LatencyCritical	[FILL]	1h	SEV1	Page immediately

Burn rate reference: 14.4× = exhausts 1-month budget in 50 hours. 6× = exhausts in ~5 days. 1× = exactly on track to exhaust in 30 days.

---

5. Error Budget Policy

Budget Remaining	Release Policy	Who Decides Exceptions
> 50%	Normal velocity	—
25–50%	Non-critical changes on hold	Engineering TL
< 25%	Feature freeze; reliability work only	Engineering TL + SRE TL
Exhausted	No changes except incident fixes	VP Engineering + SRE

Exception release process: [FILL: who approves an exception, what documentation is required, SLA for the exception decision]

---

6. Dependencies

Services or infrastructure this SLO depends on. If a dependency degrades, note whether it excludes from SLO or causes an SLO miss.

Dependency	Type	SLA Provided	Impact if Down	Excluded from SLO?
[FILL: upstream service]	synchronous	[FILL]	[FILL: availability impact]	No
[FILL: data store]	synchronous	[FILL]	[FILL: full outage]	No
[FILL: async queue]	async	[FILL]	[FILL: latency degraded]	Partial

---

7. Runbook Links

Alert	Runbook	Last Verified
[FILL: alert name]	[FILL: Confluence runbook URL]	YYYY-MM-DD
[FILL: alert name]	[FILL: Confluence runbook URL]	YYYY-MM-DD

---

8. SLO History

Track quarterly performance. Add rows at the end of each quarter.

Quarter	Availability Met?	Latency Met?	Error Budget Used	Notes
[FILL: Q1 YYYY]	Yes / No	Yes / No	[FILL: e.g. 23%]	[FILL: any notable misses or incidents]

---

9. Related Documents

• SLO Dashboard: [FILL: Grafana URL]
• Error Budget Burn Dashboard: [FILL: Grafana URL]
• Service Runbook Index: [FILL: Confluence link]
• Service Architecture Doc: [FILL: link]
• Post-mortems: [FILL: link to post-mortem folder]

---