Enterprise operations promotion to Stable requires a 30-day field pilot without data-loss
regressions. Use this gate before updating docs/feature-status.md or
docs/stable-hardening-enterprise-ops.md.
mkdir -p .spanda
date -u +%Y-%m-%d > .spanda/field-soak-start.txt
Or use the init helper:
./scripts/enterprise_ops_field_soak_init.sh
Commit the start date to your pilot branch or store it in fleet configuration management.
chmod +x scripts/field_soak_gate.sh
./scripts/field_soak_gate.sh
For the full Stable promotion workflow (soak + audit prep + smokes), see
enterprise-ops-stable-promotion.md and
./scripts/enterprise_ops_stable_promotion_gate.sh.
| Variable | Default | Purpose |
|---|---|---|
SPANDA_FIELD_SOAK_START_FILE |
.spanda/field-soak-start.txt |
UTC start date (YYYY-MM-DD) |
SPANDA_FIELD_SOAK_MIN_DAYS |
30 |
Minimum elapsed days |
scripts/enterprise_ops_smoke.sh passes.scripts/failover_drill_smoke.sh passes (when present).scripts/ota_fleet_soak.sh quick mode passes (when present).Add ./scripts/field_soak_gate.sh to your fleet promotion pipeline after the soak period. Until the
clock elapses, the script exits non-zero by design.
For Human Interaction (separate 30-day clock), use .spanda/hri-field-soak-start.txt and
./scripts/hri_stable_promotion_gate.sh — see
stable-hardening-human-interaction.md. Start the clock
with ./scripts/hri_field_soak_init.sh (one-time).
For Recovery Orchestrator (separate clock), use .spanda/recovery-field-soak-start.txt and
./scripts/recovery_orchestrator_stable_promotion_gate.sh — see
stable-hardening-recovery-orchestrator.md. Start with
./scripts/recovery_orchestrator_field_soak_init.sh (one-time).