← Back to Docs Home

SLOs and SLAs

Service Level Objectives (SLOs) and Service Level Agreements (SLAs) for QNSP Cloud services provided by CUI Labs (Pte.) Ltd.

SLOs and SLAs

Service Level Objectives (SLOs) and Service Level Agreements (SLAs) for QNSP Cloud services provided by CUI Labs (Pte.) Ltd.

Overview

  • SLOs are internal operational targets that we monitor and optimize against. They are not contractual commitments.
  • SLAs are contractual commitments with defined remedies (Service Credits) for eligible subscription tiers.

For the full SLA document with availability commitments, credit schedules, and exclusions, see Service Level Agreement.

Service Level Objectives (SLOs)

SLO targets below reflect current operational targets for QNSP Cloud hosted production. These are engineering goals, not contractual guarantees.

Availability Targets

Service SLO Target Notes
Edge Gateway 99.99% Primary ingress point
Auth Service 99.9% Authentication and token issuance
KMS Service 99.9% Key management operations
Vault Service 99.9% Secrets management
Storage Service 99.9% Encrypted object storage
Search Service 99.9% Searchable encryption
Audit Service 99.9% Audit log ingestion

Latency Targets (P99)

Operation Target Notes
Authentication < 200ms Initial login
Token refresh < 100ms Token renewal
KMS encrypt/decrypt < 100ms Single operation
Secret read < 150ms Single secret
Storage GET (< 1MB) < 300ms Small object retrieval
Storage GET (< 100MB) < 2s Medium object retrieval
Search query < 500ms SSE-X encrypted search

Error Rate Targets

Service Target
All services < 0.1% (5xx errors)

SLO Monitoring

Error Budget

The error budget represents the acceptable amount of unreliability:

error_budget = 1 - SLO_target
error_budget_remaining = error_budget - actual_error_rate

Example: For a 99.9% SLO, the monthly error budget is 0.1% (~43 minutes of downtime).

Burn Rate Alerting

Burn rate measures how quickly the error budget is being consumed:

burn_rate = actual_error_rate / (1 - SLO_target)
Burn Rate Alert Level Action
> 14.4× Critical Page on-call, immediate response
> 6× High Alert, investigate within 1 hour
> 1× Warning Monitor, investigate within 4 hours

SLO Dashboard

Real-time SLO metrics are available via API:

GET /observability/v1/slos/summary
GET /observability/v1/slos
GET /observability/v1/slos/{service}/history

Service Level Agreements (SLAs)

SLAs provide contractual availability commitments with Service Credits for eligible tiers.

SLA Eligibility by Tier

Tier SLA Coverage
Free, Dev Starter No SLA
Dev Pro, Dev Elite Monitoring only (no credits)
Business Team, Business Advanced, Business Elite 99.9% SLA with credits
Enterprise Standard, Enterprise Pro 99.95% SLA with credits
Enterprise Elite, Specialized Custom SLA per agreement

Service Credit Schedule

See the full SLA document for:

  • Detailed credit percentages
  • SLA exclusions
  • Credit request procedures
  • Measurement methodology

Incident Response

When an SLO breach occurs:

  1. Detection: Automated alerting triggers
  2. Response: On-call engineer engaged
  3. Communication: Status page updated within 15 minutes
  4. Resolution: Issue mitigated
  5. Post-Incident: Root cause analysis published
  6. Credits: Service Credits processed (if SLA applies)

For details on incident classification and response times, see Incident Response.