Files
Deep-Package-Inspection/proxy/clean_devicestat_duplicates.js
T

74 lines
2.9 KiB
JavaScript

// proxy/clean_devicestat_duplicates.js
// ─────────────────────────────────────────────────────────────────────────────
// One-time cleanup script to deduplicate historical DeviceStat records.
// Keeps only the LATEST document per (agent_uuid, ip_address) pair,
// removing all older duplicates accumulated before the upsert fix.
//
// Usage: node proxy/clean_devicestat_duplicates.js
// ─────────────────────────────────────────────────────────────────────────────
const path = require('path');
require('dotenv').config({ path: path.join(__dirname, '..', '.env.local') });
const mongoose = require('mongoose');
const MONGODB_URI = process.env.MONGODB_URI || 'mongodb://localhost:27017/backone';
const DeviceStatSchema = new mongoose.Schema({
timestamp: { type: Date },
agent_uuid: { type: String },
site_uuid: { type: String },
ip_address: { type: String },
mac_address: { type: String },
device_label: String,
device_type: String,
os_label: String,
manufacturer: String,
download: Number,
upload: Number,
flows: Number,
last_seen: String,
}, { timestamps: { createdAt: 'created_at', updatedAt: 'updated_at' } });
const DeviceStat = mongoose.model('DeviceStat', DeviceStatSchema);
async function run() {
console.log('[Cleanup] Connecting to MongoDB...');
await mongoose.connect(MONGODB_URI);
console.log('[Cleanup] Connected.');
// Find all unique (agent_uuid, ip_address) combinations
const groups = await DeviceStat.aggregate([
{ $group: {
_id: { agent_uuid: '$agent_uuid', ip_address: '$ip_address' },
ids: { $push: '$_id' },
timestamps: { $push: '$timestamp' },
count: { $sum: 1 },
}},
{ $match: { count: { $gt: 1 } } },
]);
console.log(`[Cleanup] Found ${groups.length} (agent_uuid, ip_address) pairs with duplicates.`);
let totalDeleted = 0;
for (const group of groups) {
// Sort the ids by matching timestamps - keep the latest
const paired = group.ids.map((id, i) => ({ id, ts: group.timestamps[i] }));
paired.sort((a, b) => new Date(b.ts) - new Date(a.ts));
// Keep the first (newest), delete the rest
const toDelete = paired.slice(1).map(p => p.id);
const result = await DeviceStat.deleteMany({ _id: { $in: toDelete } });
totalDeleted += result.deletedCount;
}
const remaining = await DeviceStat.countDocuments();
console.log(`[Cleanup] Done. Deleted ${totalDeleted} duplicate DeviceStat records.`);
console.log(`[Cleanup] Remaining DeviceStat documents: ${remaining}`);
await mongoose.disconnect();
}
run().catch(err => {
console.error('[Cleanup] Fatal error:', err.message);
process.exit(1);
});