Spanda

Recovery validation report

Formal validation envelope produced by the Recovery Orchestrator dry-run and execute paths.

Promotion gate

./scripts/recovery_orchestrator_stable_promotion_gate.sh

CI job: recovery-orchestrator-stable-promotion-gate in .github/workflows/ci-nightly.yml.

Field soak (30-day organizational gate):

./scripts/recovery_orchestrator_field_soak_init.sh

Sources

Layer	Validation
Assurance gates	`validate_recovery_plan`, readiness evaluation
Orchestrator	Health, capability, hardware, readiness, trust, security, mission gates
Plugin validators	Registered `[recovery.extensions]` kind `validator` require safe plans
Evidence	Immutable `OrchestratorRecoveryEvidence` records in history

CLI

spanda recovery validate rover.sd --failure gps
spanda recovery dry-run rover.sd --entity Rover --failure gps

REST

curl -s -X POST http://127.0.0.1:8787/v1/recovery/validate \
  -H 'Content-Type: application/json' \
  -d '{"failure":"gps"}'

Response includes report and validation evidence arrays.

Persistence

Evidence history is stored in control-center-recovery.json under SPANDA_CONTROL_CENTER_STATE_DIR and survives Control Center restart. Verified by recovery_history_persists_across_restart in tenant_persistence_tests.rs.

curl -s http://127.0.0.1:8787/v1/recovery/predictive
curl -s http://127.0.0.1:8787/v1/recovery/recoverable-entities
curl -s -X POST http://127.0.0.1:8787/v1/recovery/recommend \
  -H 'Content-Type: application/json' \
  -d '{"failure":"gps_loss"}'