Initial state

Dockerfile

@@ -0,0 +1,47 @@
+# Use Debian-based Python 3.13 slim image
+FROM python:3.13-slim-bookworm
+
+# Set working directory
+WORKDIR /app
+
+# Install system dependencies for rrdtool
+RUN apt-get update && apt-get install -y \
+    rrdtool \
+    librrd-dev \
+    build-essential \
+    python3-dev \
+    wget \
+    && rm -rf /var/lib/apt/lists/*
+
+# Copy requirements first for better Docker layer caching
+COPY requirements.txt .
+
+# Install Python dependencies
+RUN pip install --no-cache-dir --upgrade pip && \
+    pip install --no-cache-dir -r requirements.txt
+
+# Remove build dependencies to reduce image size
+RUN apt-get purge -y build-essential python3-dev && \
+    apt-get autoremove -y && \
+    apt-get clean
+
+# Copy application code
+COPY app/ ./app/
+
+# Create directory for RRD data
+RUN mkdir -p /app/data
+
+# Expose port
+EXPOSE 8000
+
+# Create non-root user for security
+RUN useradd --create-home --shell /bin/bash appuser && \
+    chown -R appuser:appuser /app
+USER appuser
+
+# Health check
+HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
+    CMD wget --no-verbose --tries=1 --spider http://localhost:8000/health || exit 1
+
+# Run the application
+CMD ["python", "-m", "uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000"]

app/main.py

@@ -0,0 +1,203 @@
+# app/main.py
+
+import os
+import uuid
+import json
+import logging
+from datetime import datetime
+
+from fastapi import FastAPI, Request, status
+from fastapi.responses import HTMLResponse, JSONResponse
+from fastapi.templating import Jinja2Templates
+from pydantic import BaseModel, Field, validator, constr, conlist
+from typing import Dict, List
+import uuid as uuid_lib
+
+from collections import deque
+
+from pythonjsonlogger import jsonlogger
+
+# --- Service Configuration ---
+# Generate a unique Service UUID on startup, or get it from an environment variable
+SERVICE_UUID = os.environ.get("SERVICE_UUID", str(uuid.uuid4()))
+
+# --- Logging Configuration ---
+# Get the root logger
+logger = logging.getLogger()
+logger.setLevel(logging.INFO)
+
+# Use a handler that streams to stdout
+logHandler = logging.StreamHandler()
+
+# Create a JSON formatter and add it to the handler
+# The format string adds default log attributes to the JSON output
+formatter = jsonlogger.JsonFormatter(
+    '%(asctime)s %(name)s %(levelname)s %(message)s'
+)
+logHandler.setFormatter(formatter)
+
+# Add the handler to the root logger
+# Avoid adding handlers multiple times in a uvicorn environment
+if not logger.handlers:
+    logger.addHandler(logHandler)
+
+
+class LogBuffer:
+    def __init__(self, maxlen=1000):
+        self.buffer = deque(maxlen=maxlen)
+
+    def add_log(self, record):
+        self.buffer.append({
+            'timestamp': record.get('asctime'),
+            'level': record.get('levelname'),
+            'message': record.get('message'),
+            'extra': {k: v for k, v in record.items()
+                     if k not in ['asctime', 'levelname', 'message', 'name']}
+        })
+
+    def get_logs(self, limit=100):
+        return list(self.buffer)[-limit:]
+
+# Create global log buffer
+log_buffer = LogBuffer()
+
+# Add buffer handler to logger
+buffer_handler = BufferHandler()
+logger.addHandler(buffer_handler)
+
+# Custom handler to capture logs
+class BufferHandler(logging.Handler):
+    def emit(self, record):
+        try:
+            # Format the record as JSON
+            formatter = jsonlogger.JsonFormatter()
+            log_entry = json.loads(formatter.format(record))
+            log_buffer.add_log(log_entry)
+        except Exception:
+            pass
+
+# --- FastAPI Application ---
+app = FastAPI(
+    title="Node Monitoring System",
+    description=f"A distributed monitoring system. Service UUID: {SERVICE_UUID}"
+)
+
+# Configure templates for the web interface
+templates = Jinja2Templates(directory="app/web/templates")
+
+
+# --- Data Models (as defined in the project spec) ---
+class NodeStatusModel(BaseModel):
+    uptime_seconds: int
+    load_avg: conlist(float, min_length=3, max_length=3)
+    memory_usage_percent: float
+
+class PingModel(BaseModel):
+    pings: Dict[constr(regex=r'^[0-9a-fA-F-]{36}$'), float]
+
+class StatusUpdate(BaseModel):
+    node: str = Field(..., description="Node UUID")
+    timestamp: datetime
+    status: NodeStatusModel
+    pings: Dict[str, float]
+
+    @validator('node')
+    def validate_node_uuid(cls, v):
+        try:
+            uuid_lib.UUID(v)
+            return v
+        except ValueError:
+            raise ValueError('Invalid UUID format')
+
+    @validator('pings')
+    def validate_ping_uuids(cls, v):
+        for key in v.keys():
+            try:
+                uuid_lib.UUID(key)
+            except ValueError:
+                raise ValueError(f'Invalid UUID format in pings: {key}')
+        return v
+
+
+# A mock database of known nodes for the auto-discovery demo
+# In a real app, this would be managed more dynamically
+known_nodes_db = {}
+
+
+# --- API Endpoints ---
+@app.get("/", response_class=HTMLResponse)
+async def read_root(request: Request):
+    """Serves the main web page which displays the Service UUID."""
+    logger.info(
+        "Web root accessed",
+        extra={'client_ip': request.client.host, 'service_uuid': SERVICE_UUID}
+    )
+    return templates.TemplateResponse(
+        "index.html",
+        {"request": request, "service_uuid": SERVICE_UUID}
+    )
+
+# Add the logs endpoint
+@app.get("/{service_uuid}/logs")
+async def get_logs(service_uuid: str, limit: int = 100):
+    """Get recent logs for the service."""
+    if service_uuid != SERVICE_UUID:
+        return JSONResponse(
+            status_code=404,
+            content={"error": "Service UUID not found"}
+        )
+    
+    logs = log_buffer.get_logs(limit)
+    return {
+        "service_uuid": service_uuid,
+        "log_count": len(logs),
+        "logs": logs
+    }
+
+
+@app.put("/{service_uuid}/{node_uuid}/", status_code=status.HTTP_200_OK)
+async def update_node_status(
+    service_uuid: str,
+    node_uuid: str,
+    status_update: StatusUpdate,
+    request: Request
+):
+    """Receives status updates from a node and returns a list of peers."""
+    # Log the incoming status update with structured context
+    logger.info(
+        "Received node status update",
+        extra={
+            'event_type': 'node_status_update',
+            'client_ip': request.client.host,
+            'service_uuid': service_uuid,
+            'node_uuid': node_uuid,
+            'data': status_update.dict()
+        }
+    )
+
+    # In a real implementation, you would now update the RRD database here.
+    # e.g., database.update(node_uuid, status_update)
+    # For now, we simulate the logic.
+    if service_uuid != SERVICE_UUID:
+        logger.warning(
+            "Node sent status to wrong service UUID",
+            extra={'client_node_uuid': node_uuid, 'target_uuid': service_uuid}
+        )
+        return {"error": "Service UUID mismatch", "peers": []}
+
+    # Auto-discovery logic: update our list of known nodes
+    if node_uuid not in known_nodes_db:
+        logger.info(f"New node discovered: {node_uuid}")
+    # A real app would need a strategy to handle node addresses
+    known_nodes_db[node_uuid] = {"last_seen": datetime.utcnow().isoformat(), "ip": request.client.host}
+
+    # Respond with the list of other known peers
+    peer_list = {uuid: data for uuid, data in known_nodes_db.items() if uuid != node_uuid}
+
+    return {"message": "Status received", "peers": peer_list}
+
+
+@app.get("/health")
+async def health_check():
+    """Health check endpoint for container orchestration."""
+    return {"status": "ok"}

app/web/templates/logs.html

@@ -0,0 +1,63 @@
+<!DOCTYPE html>
+<html>
+  <head>
+    <title>Node Monitor - Logs</title>
+    <style>
+body { font-family: monospace; margin: 20px; background: #1a1a1a; color: #00ff00; }
+.log-entry { margin: 5px 0; padding: 5px; border-left: 3px solid #333; }
+.log-info { border-left-color: #0066cc; }
+.log-warning { border-left-color: #ff9900; }
+.log-error { border-left-color: #cc0000; }
+.timestamp { color: #666; }
+.level { font-weight: bold; margin-right: 10px; }
+.message { color: #fff; }
+.extra { color: #888; font-size: 0.9em; margin-top: 3px; }
+h1 { color: #00ff00; }
+.refresh-info { color: #666; margin: 10px 0; }
+    </style>
+  </head>
+  <body>
+    <h1>Node Monitor Logs - {{ service_uuid }}</h1>
+    <div class="refresh-info">Auto-refreshing every 5 seconds...</div>
+
+    <div id="logs-container">
+      {% for log in logs %}
+      <div class="log-entry log-{{ log.level.lower() }}">
+        <span class="timestamp">{{ log.timestamp }}</span>
+        <span class="level">{{ log.level }}</span>
+        <span class="message">{{ log.message }}</span>
+        {% if log.extra %}
+        <div class="extra">{{ log.extra | tojson }}</div>
+        {% endif %}
+      </div>
+      {% endfor %}
+    </div>
+
+    <script>
+      // Auto-refresh logs every 5 seconds
+      setInterval(async function() {
+        try {
+          const response = await fetch('{{ service_uuid }}/logs/json');
+          const data = await response.json();
+
+          const container = document.getElementById('logs-container');
+          container.innerHTML = '';
+
+          data.logs.forEach(log => {
+            const div = document.createElement('div');
+            div.className = `log-entry log-${log.level.toLowerCase()}`;
+            div.innerHTML = `
+                          <span class="timestamp">${log.timestamp}</span>
+                          <span class="level">${log.level}</span>
+                          <span class="message">${log.message}</span>
+                          ${log.extra ? `<div class="extra">${JSON.stringify(log.extra)}</div>` : ''}
+                          `;
+            container.appendChild(div);
+          });
+        } catch (error) {
+          console.error('Failed to refresh logs:', error);
+        }
+      }, 5000);
+    </script>
+  </body>
+</html>

requirements.txt

@@ -0,0 +1,6 @@
+fastapi==0.104.1
+uvicorn[standard]==0.24.0
+python-rrdtool==1.4.7
+jinja2==3.1.2
+python-multipart==0.0.6
+python-json-logger==2.0.7