Skip to main content

Gate5 SLO And Alerting Baseline

Objective​

Define one explicit, versioned reliability baseline for critical Gate5 runtime paths and enforce it in CI.

Source Of Truth​

  1. Baseline version: gate5-slo-alert-baseline.v1
  2. Versioned config: contracts/config/slo-alert-baseline.v1.json
  3. Verification gate: contracts/scripts/check-slo-alert-baseline.py
  4. CI enforcement path: .github/workflows/contracts-governance.yml

SLO Matrix​

SLO IDServiceCritical PathSLITargetWindowOwner
platform_api_request_success_rateplatform-apiexecution,risk,research,reconciliation,conversation2xx_or_contractual_4xx_rate>=99.5%30dTeam F
risk_pretrade_latency_p95_msrisk-pretraderisk,executionp95_latency_ms<=2007dTeam F
research_market_scan_latency_p95_msresearchresearchp95_latency_ms<=25007dTeam F
reconciliation_cycle_success_ratereconciliationreconciliationsuccessful_cycle_rate>=99.0%30dTeam F
conversation_turn_latency_p95_msconversationconversationp95_latency_ms<=12007dTeam F

Alert Matrix​

Alert IDSLO IDConditionSeverityOwner
alert_platform_api_success_rate_burnplatform_api_request_success_rateerror_budget_burn_2h>=5%SEV-2Team F
alert_risk_pretrade_latencyrisk_pretrade_latency_p95_msp95_latency_ms>200_for_15mSEV-2Team F
alert_research_latencyresearch_market_scan_latency_p95_msp95_latency_ms>2500_for_15mSEV-3Team F
alert_reconciliation_success_ratereconciliation_cycle_success_ratesuccessful_cycle_rate<99.0_for_15mSEV-2Team F
alert_conversation_turn_latencyconversation_turn_latency_p95_msp95_latency_ms>1200_for_15mSEV-3Team F

Governance Rules​

  1. Any SLO or alert semantics change must update both this portal page and contracts/config/slo-alert-baseline.v1.json in the same PR.
  2. contracts/scripts/check-slo-alert-baseline.py must pass in CI before merge.
  3. Missing owner, target, or ID alignment is a release-gating failure for Gate5.

Traceability​

  • Reliability parent: #81
  • Gate5 reliability closure issue: #45
  • Gate workflow template: docs/portal/operations/gate-workflow.md