Getting Started

# Add the Chaos Mesh repository to the Helm repository
helm repo add chaos-mesh https://charts.chaos-mesh.org

# To see charts that can be installed
helm search repo chaos-mesh

# Install Chaos Mesh under the chaos-mesh namespace
kubectl create ns chaos-mesh

helm install chaos-mesh chaos-mesh/chaos-mesh -n=chaos-mesh
    --set chaosDaemon.runtime=containerd \
    --set chaosDaemon.socketPath=/run/containerd/containerd.sock \
    --version 2.7.2

kubectl get svc -n chaos-mesh
NAME                            TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)                                 AGE
chaos-daemon                    ClusterIP   None            <none>        31767/TCP,31766/TCP                     30m
chaos-dashboard                 NodePort    10.97.177.177   <none>        2333:31632/TCP,2334:32683/TCP           30m
chaos-mesh-controller-manager   ClusterIP   10.105.196.18   <none>        443/TCP,10081/TCP,10082/TCP,10080/TCP   30m
chaos-mesh-dns-server           ClusterIP   10.96.127.183   <none>        53/UDP,53/TCP,9153/TCP,9288/TCP         30m

kind: ServiceAccount
apiVersion: v1
metadata:
  namespace: default
  name: account-cluster

---
kind: ClusterRole
apiVersion: rbac.authorization.k8s.io/v1
metadata:
  name: role-cluster
rules:
  - apiGroups: [""]
    resources: ["pods", "namespaces"]
    verbs: ["get", "watch", "list"]
  - apiGroups: ["chaos-mesh.org"]
    resources: ["*"]
    verbs: ["get", "list", "watch", "create", "delete", "patch", "update"]

---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: bind-cluster
subjects:
  - kind: ServiceAccount
    name: account-cluster
    namespace: default
roleRef:
  kind: ClusterRole
  name: role-cluster
  apiGroup: rbac.authorization.k8s.io

kubectl apply -f rbac.yaml

kubectl create token account-cluster -n default

# Create a namespace for the benchmark application
kubectl create namespace ts0

# Deploy the train-ticket application
# ts is the prefix for the namespace and 1 is the number of pedestal instances
make install-releases PEDESTAL_KEY=ts PEDESTAL_COUNT=1

git clone https://github.com/OperationsPAI/AegisLab.git
cd AegisLab

make install-secrets

make install-hostpath

# Start all services with Docker Compose
make local-debug

# This will start:
# - MySQL database
# - Redis cache
# - Jaeger tracing
# - BuildKit daemon
# - AegisLab API server

# Check if all services are running
docker ps

# Test API access
curl http://localhost:8082/system/health

# Access Swagger documentation
open http://localhost:8082/swagger/index.html

# Install RCABench SDK
pip install rcabench

# Or install from source
cd sdk/python
pip install -e .

from rcabench import RCABenchSDK
import time

# Initialize SDK
sdk = RCABenchSDK("http://localhost:8082")

# Verify connection
health = sdk.health_check()
print(f"RCABench status: {health}")

# List available algorithms
algorithms = sdk.algorithm.list()
print("Available algorithms:")
for algo in algorithms:
    print(f"  - {algo['name']}: {algo['description']}")

# List available datasets
datasets = sdk.dataset.list()
print("Available datasets:")
for dataset in datasets:
    print(f"  - {dataset['name']}: {dataset['description']}")

# Define fault injection request
fault_request = DtoSubmitInjectionReq(
    algorithms=[
        DtoAlgorithmItem(name="traceback"),
    ], # Algorithm execution container name
    benchmark="clickhouse", # Data collection container name
    container_name="ts_cn", # Pedestal container name
    container_tag="v1.0.0-213-gf9294111", # Pedestal container tag
    interval=2, # The whole period in minutes
    project_name="pair_diagnosis",
    pre_duration=1, # Normal time in minutes
    specs=[
        HandlerNode(
            children={
                "4": HandlerNode(
                    children={
                        "0": HandlerNode(value=1), # Duration in minutes
                        "1": HandlerNode(value=0), # Namespace prefix ts is to 0
                        "2": HandlerNode(value=1), # Container Index
                        "3": HandlerNode(value=1), # CPU Load Percentage
                        "4": HandlerNode(value=1), # CPU Stress Threads
                    },
                )
            },
            value=4,
        )
    ],
    labels=[
        DtoLabelItem(key="env", value=env),
        DtoLabelItem(key="batch", value="bootstrap"),
    ],
)

# Execute fault injection
injection_result = sdk.injection.execute(fault_request)