Healing API

API endpoints for the self-healing pipeline — ledger, trust states, and suggestions.

Endpoints

GET /api/healing/ledger

Returns recent heal outcomes.

Auth: Any authenticated user

Query Parameters:

Parameter	Type	Default	Description
limit	int	50	Max results to return
rule_id	string	—	Filter by rule ID
target	string	—	Filter by target (hostname/container)

Response: Array of outcome objects

[
  {
    "id": 1,
    "correlation_key": "nginx:cpu_high",
    "rule_id": "cpu_high",
    "condition": "cpu_percent > 90",
    "target": "nginx",
    "action_type": "restart_container",
    "action_params": "{\"container\": \"nginx\"}",
    "escalation_step": 0,
    "action_success": 1,
    "verified": 1,
    "duration_s": 12.5,
    "trust_level": "execute",
    "source": "alert",
    "created_at": 1711100000
  }
]

GET /api/healing/effectiveness

Returns success rate for a specific action/condition pair.

Auth: Any authenticated user

Query Parameters (required):

Parameter	Type	Description
action_type	string	Action type (e.g., restart_container)
condition	string	Condition string (e.g., cpu_percent > 90)
target	string	Optional target filter

Response:

{
  "action_type": "restart_container",
  "condition": "cpu_percent > 90",
  "success_rate": 87.5
}

GET /api/healing/suggestions

Returns active (non-dismissed) suggestions.

Auth: Any authenticated user

Response: Array of suggestion objects

[
  {
    "id": 1,
    "category": "recurring_issue",
    "severity": "warning",
    "message": "Target 'nginx' healed 15 times for rule 'cpu_high'.",
    "rule_id": "cpu_high",
    "dismissed": 0,
    "created_at": 1711100000
  }
]

POST /api/healing/suggestions/{suggestion_id}/dismiss

Dismiss a suggestion.

Auth: Operator or Admin

Response:

{ "success": true }

GET /api/healing/trust

Returns trust state for all rules.

Auth: Any authenticated user

Response: Array of trust state objects

[
  {
    "rule_id": "cpu_high",
    "current_level": "approve",
    "ceiling": "execute",
    "promoted_at": 1711100000,
    "demoted_at": null,
    "promotion_count": 1,
    "demotion_count": 0,
    "last_evaluated": 1711100000
  }
]

POST /api/healing/trust/{rule_id}/promote

Manually promote a rule's trust level.

Auth: Admin only

Body:

{ "level": "approve" }

level must be "approve" or "execute".

Response:

{
  "success": true,
  "rule_id": "cpu_high",
  "new_level": "approve"
}