File size: 4,567 Bytes
d02897f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 | # Copyright (c) Meta Platforms, Inc.
"""
Pydantic models for OpenOps environment
"""
from typing import Dict, List, Optional
from pydantic import BaseModel, Field
class IncidentAction(BaseModel):
"""
Action taken by the agent.
Represents a single action in the incident management workflow.
"""
action_id: int = Field(..., ge=0, le=20, description="Action ID (0-20)")
task_id: int = Field(default=1, ge=1, le=3, description="Task ID (1=easy, 2=medium, 3=hard)")
class Config:
json_schema_extra = {
"example": {
"action_id": 0,
"task_id": 1
}
}
class IncidentObservation(BaseModel):
"""
Observation returned to agent after each step.
Contains partial information about the system state (investigation reveals more).
"""
active_alerts: List[str] = Field(
default_factory=list,
description="List of active system alerts"
)
service_status: Dict[str, str] = Field(
default_factory=dict,
description="Status of each service (healthy/degraded/down)"
)
recent_logs: Dict[str, List[str]] = Field(
default_factory=dict,
description="Logs from inspected services only"
)
metrics_summary: Dict[str, Dict[str, float]] = Field(
default_factory=dict,
description="Metrics for checked services (CPU, memory, latency)"
)
customer_complaints: int = Field(
default=0,
description="Number of customer complaints received"
)
time_elapsed: int = Field(
default=0,
description="Minutes since incident started"
)
revenue_loss: float = Field(
default=0.0,
description="Estimated revenue loss in USD"
)
teams_notified: bool = Field(
default=False,
description="Whether engineering team has been notified"
)
status_page_updated: bool = Field(
default=False,
description="Whether public status page has been updated"
)
reward: float = Field(
default=0.0,
description="Reward received for this step"
)
done: bool = Field(
default=False,
description="Whether episode is complete"
)
class Config:
json_schema_extra = {
"example": {
"active_alerts": ["CRITICAL: API service down"],
"service_status": {
"api": "down",
"database": "healthy"
},
"recent_logs": {
"api": ["ERROR: Out of memory"]
},
"customer_complaints": 45,
"time_elapsed": 5,
"revenue_loss": 5000.0,
"teams_notified": False,
"status_page_updated": False,
"reward": 0.05,
"done": False
}
}
class IncidentState(BaseModel):
"""
Internal environment state (hidden from agent).
Contains ground truth about the incident for evaluation.
"""
task_id: int = Field(..., ge=1, le=3, description="Task difficulty level")
incident_type: str = Field(..., description="Type of incident")
affected_services: List[str] = Field(
default_factory=list,
description="Services affected by the incident"
)
root_cause: str = Field(..., description="Root cause of the incident")
service_status: Dict[str, str] = Field(
default_factory=dict,
description="Current status of all services"
)
correct_mitigation: List[str] = Field(
default_factory=list,
description="Correct mitigation actions for this incident"
)
revenue_loss: float = Field(
default=0.0,
description="Accumulated revenue loss"
)
customer_complaints: int = Field(
default=0,
description="Accumulated customer complaints"
)
class Config:
json_schema_extra = {
"example": {
"task_id": 1,
"incident_type": "api_crash",
"affected_services": ["api"],
"root_cause": "out_of_memory",
"service_status": {
"api": "down",
"database": "healthy",
"auth": "healthy",
"frontend": "degraded"
},
"correct_mitigation": ["restart_api"],
"revenue_loss": 0.0,
"customer_complaints": 0
}
} |