server/tasks/k8s_pod.py

"""Task: Kubernetes Pod Failures — MEDIUM.

Agent fixes common pod failure scenarios:
OOMKilled, ImagePullBackOff, wrong command, missing ConfigMap, liveness probe.
"""

from server.models import TaskDifficulty
from server.tasks.base import BaseTask


class K8sPodTask(BaseTask):
    NAME = "Kubernetes Pod Failures"
    DESCRIPTION = "Fix Kubernetes pod failures including CrashLoopBackOff, ImagePullBackOff, and resource issues"
    DIFFICULTY = TaskDifficulty.MEDIUM
    AVAILABLE_SECRETS = []

    SCENARIOS = [
        # Scenario 1: CrashLoopBackOff — OOMKilled (memory limit too low)
        {
            "id": "oom_killed",
            "files": [
                {
                    "path": "k8s/deployment.yaml",
                    "type": "kubernetes",
                    "content": (
                        "apiVersion: apps/v1\n"
                        "kind: Deployment\n"
                        "metadata:\n"
                        "  name: api-server\n"
                        "spec:\n"
                        "  replicas: 3\n"
                        "  selector:\n"
                        "    matchLabels:\n"
                        "      app: api\n"
                        "  template:\n"
                        "    metadata:\n"
                        "      labels:\n"
                        "        app: api\n"
                        "    spec:\n"
                        "      containers:\n"
                        "      - name: api\n"
                        '        image: myapp:v1.2.3\n'
                        "        resources:\n"
                        "          limits:\n"
                        '            memory: "64Mi"\n'
                        '            cpu: "100m"\n'
                        "        ports:\n"
                        "        - containerPort: 8080\n"
                    ),
                }
            ],
            "error": {
                "phase": "k8s_runtime",
                "message": (
                    "$ kubectl get pods\n"
                    "NAME                          READY   STATUS             RESTARTS   AGE\n"
                    "api-server-7d4b8c9f5-x2k9m   0/1     CrashLoopBackOff   5          3m\n"
                    "\n"
                    "$ kubectl describe pod api-server-7d4b8c9f5-x2k9m\n"
                    "...\n"
                    "State:          Waiting\n"
                    "  Reason:       CrashLoopBackOff\n"
                    "Last State:     Terminated\n"
                    "  Reason:       OOMKilled\n"
                    "  Exit Code:    137\n"
                    "...\n"
                    "Events:\n"
                    "  Warning  OOMKilling  3m  kubelet  Memory limit 64Mi exceeded"
                ),
            },
            "expected_fixes": [
                {
                    "file": "k8s/deployment.yaml",
                    "type": "contains",
                    "expected": 'memory: "256Mi"',
                    "hint": "Container is OOMKilled with 64Mi limit. The app needs at least 256Mi.",
                }
            ],
        },

        # Scenario 2: ImagePullBackOff — image tag typo
        {
            "id": "image_pull_backoff",
            "files": [
                {
                    "path": "k8s/deployment.yaml",
                    "type": "kubernetes",
                    "content": (
                        "apiVersion: apps/v1\n"
                        "kind: Deployment\n"
                        "metadata:\n"
                        "  name: web-app\n"
                        "spec:\n"
                        "  replicas: 2\n"
                        "  selector:\n"
                        "    matchLabels:\n"
                        "      app: web\n"
                        "  template:\n"
                        "    metadata:\n"
                        "      labels:\n"
                        "        app: web\n"
                        "    spec:\n"
                        "      containers:\n"
                        "      - name: web\n"
                        "        image: nginx:latset\n"
                        "        ports:\n"
                        "        - containerPort: 80\n"
                    ),
                }
            ],
            "error": {
                "phase": "k8s_runtime",
                "message": (
                    "$ kubectl get pods\n"
                    "NAME                       READY   STATUS             RESTARTS   AGE\n"
                    "web-app-5f8d7b6c4-abc12    0/1     ImagePullBackOff   0          2m\n"
                    "\n"
                    "$ kubectl describe pod web-app-5f8d7b6c4-abc12\n"
                    "...\n"
                    "Events:\n"
                    '  Warning  Failed   2m   kubelet  Failed to pull image "nginx:latset": '
                    "rpc error: code = NotFound desc = failed to pull and unpack image: "
                    "reference not found\n"
                    "  Warning  Failed   2m   kubelet  Error: ImagePullBackOff\n"
                    "..."
                ),
            },
            "expected_fixes": [
                {
                    "file": "k8s/deployment.yaml",
                    "type": "contains",
                    "expected": "image: nginx:latest",
                    "hint": "Image tag has a typo: 'latset' should be 'latest'",
                }
            ],
        },

        # Scenario 3: CrashLoopBackOff — wrong command
        {
            "id": "wrong_command",
            "files": [
                {
                    "path": "k8s/deployment.yaml",
                    "type": "kubernetes",
                    "content": (
                        "apiVersion: apps/v1\n"
                        "kind: Deployment\n"
                        "metadata:\n"
                        "  name: worker\n"
                        "spec:\n"
                        "  replicas: 1\n"
                        "  selector:\n"
                        "    matchLabels:\n"
                        "      app: worker\n"
                        "  template:\n"
                        "    metadata:\n"
                        "      labels:\n"
                        "        app: worker\n"
                        "    spec:\n"
                        "      containers:\n"
                        "      - name: worker\n"
                        "        image: python:3.11-slim\n"
                        "        command: [\"python\", \"workers.py\"]\n"
                        "        resources:\n"
                        "          limits:\n"
                        '            memory: "512Mi"\n'
                        '            cpu: "500m"\n'
                    ),
                },
                {
                    "path": "app/worker.py",
                    "type": "other",
                    "content": (
                        "import time\n"
                        "\n"
                        "def main():\n"
                        "    while True:\n"
                        "        print('Processing...')\n"
                        "        time.sleep(5)\n"
                        "\n"
                        "if __name__ == '__main__':\n"
                        "    main()\n"
                    ),
                },
            ],
            "error": {
                "phase": "k8s_runtime",
                "message": (
                    "$ kubectl get pods\n"
                    "NAME                      READY   STATUS             RESTARTS   AGE\n"
                    "worker-6b8f9d7c4-kj3m2    0/1     CrashLoopBackOff   4          2m\n"
                    "\n"
                    "$ kubectl logs worker-6b8f9d7c4-kj3m2\n"
                    "python: can't open file '/workers.py': [Errno 2] No such file or directory\n"
                    "\n"
                    "$ kubectl describe pod worker-6b8f9d7c4-kj3m2\n"
                    "...\n"
                    "State:          Waiting\n"
                    "  Reason:       CrashLoopBackOff\n"
                    "Last State:     Terminated\n"
                    "  Reason:       Error\n"
                    "  Exit Code:    2\n"
                    "..."
                ),
            },
            "expected_fixes": [
                {
                    "file": "k8s/deployment.yaml",
                    "type": "contains",
                    "expected": 'command: ["python", "worker.py"]',
                    "hint": "The command references 'workers.py' but the file is named 'worker.py' (no 's')",
                }
            ],
        },

        # Scenario 4: CreateContainerConfigError — missing ConfigMap
        {
            "id": "missing_configmap",
            "files": [
                {
                    "path": "k8s/deployment.yaml",
                    "type": "kubernetes",
                    "content": (
                        "apiVersion: apps/v1\n"
                        "kind: Deployment\n"
                        "metadata:\n"
                        "  name: backend\n"
                        "spec:\n"
                        "  replicas: 2\n"
                        "  selector:\n"
                        "    matchLabels:\n"
                        "      app: backend\n"
                        "  template:\n"
                        "    metadata:\n"
                        "      labels:\n"
                        "        app: backend\n"
                        "    spec:\n"
                        "      containers:\n"
                        "      - name: backend\n"
                        "        image: mybackend:v2.0\n"
                        "        ports:\n"
                        "        - containerPort: 8080\n"
                        "        envFrom:\n"
                        "        - configMapRef:\n"
                        "            name: app-config\n"
                        "        resources:\n"
                        "          limits:\n"
                        '            memory: "512Mi"\n'
                        '            cpu: "500m"\n'
                    ),
                },
            ],
            "error": {
                "phase": "k8s_runtime",
                "message": (
                    "$ kubectl get pods\n"
                    "NAME                       READY   STATUS                       RESTARTS   AGE\n"
                    "backend-5c9d8f7b6-lm4n5    0/1     CreateContainerConfigError   0          1m\n"
                    "\n"
                    "$ kubectl describe pod backend-5c9d8f7b6-lm4n5\n"
                    "...\n"
                    "Events:\n"
                    '  Warning  Failed  1m  kubelet  Error: configmap "app-config" not found\n'
                    "..."
                ),
            },
            "expected_fixes": [
                {
                    "file": "k8s/configmap.yaml",
                    "type": "contains",
                    "expected": "name: app-config",
                    "hint": "The ConfigMap 'app-config' is referenced but doesn't exist. Create a ConfigMap manifest.",
                }
            ],
        },

        # Scenario 5: Pod not ready — liveness probe failing
        {
            "id": "liveness_probe_failing",
            "files": [
                {
                    "path": "k8s/deployment.yaml",
                    "type": "kubernetes",
                    "content": (
                        "apiVersion: apps/v1\n"
                        "kind: Deployment\n"
                        "metadata:\n"
                        "  name: api\n"
                        "spec:\n"
                        "  replicas: 2\n"
                        "  selector:\n"
                        "    matchLabels:\n"
                        "      app: api\n"
                        "  template:\n"
                        "    metadata:\n"
                        "      labels:\n"
                        "        app: api\n"
                        "    spec:\n"
                        "      containers:\n"
                        "      - name: api\n"
                        "        image: myapi:v3.1\n"
                        "        ports:\n"
                        "        - containerPort: 8080\n"
                        "        livenessProbe:\n"
                        "          httpGet:\n"
                        "            path: /healthz\n"
                        "            port: 3000\n"
                        "          initialDelaySeconds: 5\n"
                        "          periodSeconds: 10\n"
                        "        readinessProbe:\n"
                        "          httpGet:\n"
                        "            path: /ready\n"
                        "            port: 8080\n"
                        "          initialDelaySeconds: 5\n"
                        "          periodSeconds: 10\n"
                        "        resources:\n"
                        "          limits:\n"
                        '            memory: "512Mi"\n'
                        '            cpu: "500m"\n'
                    ),
                },
            ],
            "error": {
                "phase": "k8s_runtime",
                "message": (
                    "$ kubectl get pods\n"
                    "NAME                   READY   STATUS    RESTARTS      AGE\n"
                    "api-7f8d9c6b5-gh7j8    0/1     Running   3 (30s ago)  2m\n"
                    "\n"
                    "$ kubectl describe pod api-7f8d9c6b5-gh7j8\n"
                    "...\n"
                    "Events:\n"
                    "  Warning  Unhealthy  90s  kubelet  Liveness probe failed: "
                    "Get \"http://10.244.0.5:3000/healthz\": dial tcp 10.244.0.5:3000: "
                    "connect: connection refused\n"
                    "  Normal   Killing    90s  kubelet  Container api failed liveness probe, "
                    "will be restarted\n"
                    "...\n"
                    "\n"
                    "Note: The application listens on port 8080, not 3000."
                ),
            },
            "expected_fixes": [
                {
                    "file": "k8s/deployment.yaml",
                    "type": "contains",
                    "expected": "port: 8080\n          initialDelaySeconds: 5\n          periodSeconds: 10\n        readinessProbe:",
                    "hint": "The liveness probe port (3000) doesn't match the container port (8080). Change liveness probe port to 8080.",
                }
            ],
        },
    ]