Policies

Governance rules that gate automated remediation. RPL policies set confidence thresholds, blast-radius caps, and approval routing. Every workflow trigger evaluates the active policies before any executor mutation runs — this is the surface that enforces Obvia's "AI assists, never decides" contract.

Total
0
Active
0
Draft
0
Triggers covered
0

No active policies

Every workflow currently triggers without an RPL gate. That is acceptable for the pilot scaffolding but blocks the G2 execution gate. Seed at least one approval-required policy per remediation tier before enabling auto-execution.

Example RPL policy 

policy "production-restart-requires-approval" {
  scope = "service"
  scope_value = "payment-service"
  trigger = "remediation.execute"

  rule "low-confidence-needs-human" {
    when    confidence < 0.85
    action  require_approval
    routes  team:platform-oncall
  }

  rule "high-blast-radius-blocked" {
    when    blast_radius_level == "critical"
    action  deny
  }
}

Manage via API (visual editor lands post-pilot) 

# List policies
curl -H "X-Org-Id: $ORG_ID" \
  $OBVIA_API/v1/policies

# Activate a policy
curl -X POST -H "X-Org-Id: $ORG_ID" \
  $OBVIA_API/v1/policies/$POLICY_ID/activate