Datasets:

Shravan2020
/

acea

Modalities:

Text

Formats:

text

Size:

< 1K

Libraries:

Datasets

Dataset card Data Studio Files Files and versions

xet

Community

Dataset Viewer

Auto-converted to Parquet Duplicate

Split (1)

train · 438 rows

text stringlengths 0 101
FROM python:3.11-slim

# Hugging Face Spaces compatibility
ENV HOME=/home/user
ENV PATH=/home/user/.local/bin:$PATH
WORKDIR /app

RUN apt-get update && apt-get install -y --no-install-recommends \
curl \
&& rm -rf /var/lib/apt/lists/*

COPY acea/requirements.txt ./requirements.txt
RUN pip install --no-cache-dir -r requirements.txt

COPY acea/ .

RUN touch backend/__init__.py env/__init__.py

EXPOSE 7860

HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
CMD curl -f http://localhost:7860/health \|\| exit 1

CMD ["uvicorn", "backend.main:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "1"]
fastapi==0.115.0
uvicorn[standard]==0.30.6
pydantic==2.8.2
groq==0.9.0
httpx==0.27.2
python-multipart==0.0.12
name: autonomous-enterprise-chaosops-arena
version: "1.0.0"
description: >
AI training environment simulating an enterprise DevOps + Incident Response
engineer operating under dynamic, chaotic conditions. Agents must triage
incidents, diagnose failures, and stabilize production systems under pressure.

author: ACEA Team
license: MIT

environment:
class: ACEAEnvironment
module: env.environment
api_version: "openenv-1.0"

api:
reset:
description: Reset environment to initial state for given difficulty
returns: Observation
step:
description: Execute one action step
input: Action
returns:
observation: Observation
reward: float
done: bool
info: dict
state:
description: Return full internal environment state
returns: dict

observation_space:
type: structured
schema: env.models.Observation
fields:
alerts:
type: list[Alert]
description: Active system alerts with severity and service attribution
logs:
type: list[LogEntry]
description: Recent system logs (AWS/Kubernetes style)
tickets:
type: list[Ticket]
description: Customer support tickets including escalations
system_health:
type: SystemHealth
description: Real-time system metrics (CPU, memory, latency, uptime, error_rate)
active_incidents:
type: list[Incident]
description: Currently active incidents requiring resolution
risk_level:
type: enum[low, medium, high, critical]
description: Overall system risk classification
time_elapsed:
type: int
description: Seconds elapsed since episode start
chaos_events:
type: list[ChaosEvent]
description: Chaos events injected this step
step_count:
type: int
description: Current step index

action_space:
type: structured
schema: env.models.Action
fields:
type:
type: enum
values:

End of preview. Expand in Data Studio

🔥 Autonomous Enterprise ChaosOps Arena (ACEA)

An OpenEnv-compatible AI training environment where agents act as enterprise SRE engineers navigating real-world infrastructure chaos.

🎯 Environment Description

ACEA simulates a production enterprise environment where an AI agent must act as an incident response engineer. The agent receives live alerts, logs, tickets, and system health metrics, and must take actions to resolve incidents under dynamic, escalating chaos conditions.

This is NOT a game or toy — it models real enterprise SRE scenarios including cascading failures, security breaches, and executive escalations.

🧠 Observation Space

Field	Type	Description
`alerts`	`List[Alert]`	Active system alerts with severity
`logs`	`List[LogEntry]`	AWS/Kubernetes-style log entries
`tickets`	`List[Ticket]`	Customer support tickets with escalation
`system_health`	`SystemHealth`	CPU, memory, latency, uptime, error rate
`active_incidents`	`List[Incident]`	Unresolved incidents requiring action
`risk_level`	`enum`	`low / medium / high / critical`
`chaos_events`	`List[ChaosEvent]`	Injected chaos events this step
`time_elapsed`	`int`	Seconds since episode start

⚡ Action Space

{
  "type": "isolate_system",
  "target": "auth-service",
  "priority": 4,
  "reasoning": "Security breach detected — isolating to prevent lateral movement.",
  "parameters": {}
}

Action types: restart_service, scale_system, debug_issue, notify_user, ignore, isolate_system

🎮 Tasks (Easy → Medium → Hard)

Task	Difficulty	Incidents	Chaos	Max Steps
Single Service Degradation	🟢 Easy	1	0.4x	10
Multi-Service Cascade Failure	🟡 Medium	2	1.0x	15
Triple-Threat Crisis	🔴 Hard	4	2.2x	20

🧮 Reward Function (0.0 → 1.0)

Downloads last month: 9