Start here
Error-budget calculations, balances and burn rates
Connect the objective, measurement window and eligible population.
Read the articleSLOs & reliability
Connect service objectives to error budgets, alerting and release decisions.
Work through the calculation
Error-budget template: calculations and release-policy fields
Use the calculation examples and policy fields for a service record.
Read the guideAgree on the decision
What an exhausted error budget changes
Define the release policy before an incident puts it under pressure.
Read the articleReview the release gate
Release gates that hold up under incident pressure
Connect policy thresholds to an owned operational decision.
Read the guidePut it into practice
Calculator
Error-budget calculator
Calculate the failure allowance and remaining budget in event counts or minutes.
- Input
- An SLO and eligible event or time totals
- Output
- Allowed, consumed and remaining budget in the chosen units
Calculator
SLO burn-rate calculator
Compare two windows of error rates with the SLO and your chosen burn threshold.
- Input
- An SLO and counts for two nested windows
- Output
- Two-window burn rates with threshold and scope checks
Related AI news
News ·
What Nano Banana 2.1 changes in an image-generation service
Google's new image model rejects familiar generation settings. Follow a thumbnail job through its outgoing request, retries and delivery to evaluate the upgrade.
More on slos & reliability
Guide · 9 min read
Measuring streaming LLM latency from first text to completed work
A client-side SSE probe separates first answer text from normal completion. Fixed arrivals and outcome counts show whether a serving change delivers more useful answers.
Guide · 8 min read
Measuring a latency SLO with Prometheus histograms
Build a latency SLI from Prometheus histogram buckets, include failed requests, and compare classic and native queries with a worked search-service example.
Guide · 8 min read
Why a Prometheus error-rate query can return no data
Zero errors, no traffic and absent series produce different PromQL results. Synthetic tests show how ratios, fallback zeros and absence checks behave for a checkout service.
Guide · 7 min read
Kubernetes disruption budgets during GPU maintenance
A disruption budget limits unavailable pods while an inference replica is replaced. GPU placement, model warmup and surviving capacity determine how requests fare during that interval.