MongoDB Replica Sets: High Availability and Failover with SQL-Style Database Operations

-- Single database server limitations
-- Single point of failure scenarios:

-- Hardware failure
SELECT order_id, customer_id, total_amount 
FROM orders
WHERE status = 'pending';
-- ERROR: Connection failed - server hardware malfunction

-- Network partition
UPDATE inventory 
SET quantity = quantity - 5 
WHERE product_id = 'LAPTOP001';
-- ERROR: Network timeout - server unreachable

-- Planned maintenance
ALTER TABLE users ADD COLUMN preferences JSONB;
-- ERROR: Database offline for maintenance

-- Data corruption
SELECT * FROM critical_business_data;
-- ERROR: Table corrupted, data unreadable

// MongoDB replica set provides automatic failover
const replicaSetConnection = {
  hosts: [
    'mongodb-primary.example.com:27017',
    'mongodb-secondary1.example.com:27017', 
    'mongodb-secondary2.example.com:27017'
  ],
  replicaSet: 'production-rs',
  readPreference: 'primaryPreferred',
  writeConcern: { w: 'majority', j: true }
};

// Automatic failover handling
db.orders.insertOne({
  customer_id: ObjectId("64f1a2c4567890abcdef1234"),
  items: [{ product: 'laptop', quantity: 1, price: 1200 }],
  total_amount: 1200,
  status: 'pending',
  created_at: new Date()
});
// Automatically routes to available primary server
// Fails over seamlessly if primary becomes unavailable

// Replica set topology
{
  "_id": "production-rs",
  "version": 1,
  "members": [
    {
      "_id": 0,
      "host": "mongodb-primary.example.com:27017",
      "priority": 2,      // Higher priority = preferred primary
      "votes": 1,         // Participates in elections
      "arbiterOnly": false
    },
    {
      "_id": 1, 
      "host": "mongodb-secondary1.example.com:27017",
      "priority": 1,      // Can become primary
      "votes": 1,
      "arbiterOnly": false,
      "hidden": false     // Visible to clients
    },
    {
      "_id": 2,
      "host": "mongodb-secondary2.example.com:27017", 
      "priority": 1,
      "votes": 1,
      "arbiterOnly": false,
      "buildIndexes": true,
      "tags": { "datacenter": "west", "usage": "analytics" }
    },
    {
      "_id": 3,
      "host": "mongodb-arbiter.example.com:27017",
      "priority": 0,      // Cannot become primary
      "votes": 1,         // Votes in elections only
      "arbiterOnly": true // No data storage
    }
  ],
  "settings": {
    "electionTimeoutMillis": 10000,
    "heartbeatIntervalMillis": 2000,
    "catchUpTimeoutMillis": 60000
  }
}

-- Conceptual SQL cluster configuration
CREATE CLUSTER production_cluster AS (
  -- Primary database server
  PRIMARY SERVER db1.example.com 
    WITH PRIORITY = 2,
         VOTES = 1,
         AUTO_FAILOVER = TRUE,

  -- Secondary servers for redundancy
  SECONDARY SERVER db2.example.com
    WITH PRIORITY = 1,
         VOTES = 1,
         READ_ALLOWED = TRUE,
         REPLICATION_ROLE = 'synchronous',

  SECONDARY SERVER db3.example.com  
    WITH PRIORITY = 1,
         VOTES = 1,
         READ_ALLOWED = TRUE,
         REPLICATION_ROLE = 'asynchronous',
         DATACENTER = 'west',

  -- Witness server for quorum
  WITNESS SERVER witness.example.com
    WITH VOTES = 1,
         DATA_STORAGE = FALSE,
         ELECTION_ONLY = TRUE
)
WITH ELECTION_TIMEOUT = 10000ms,
     HEARTBEAT_INTERVAL = 2000ms,
     FAILOVER_MODE = 'automatic';

// Oplog (operations log) structure
{
  "ts": Timestamp(1693547204, 1),
  "t": NumberLong("1"),
  "h": NumberLong("4321"),
  "v": 2,
  "op": "i",  // operation type: i=insert, u=update, d=delete
  "ns": "ecommerce.orders",
  "o": {  // operation document
    "_id": ObjectId("64f1a2c4567890abcdef1234"),
    "customer_id": ObjectId("64f1a2c4567890abcdef5678"),
    "total_amount": 159.98,
    "status": "pending"
  }
}

// Replication flow:
// 1. Write operation executed on primary
// 2. Operation recorded in primary's oplog
// 3. Secondary servers read and apply oplog entries
// 4. Write acknowledged based on write concern

// 1. Start MongoDB instances with replica set configuration
// Primary server (db1.example.com)
mongod --replSet production-rs --dbpath /data/db --logpath /var/log/mongodb.log --fork --bind_ip 0.0.0.0

// Secondary servers (db2.example.com, db3.example.com)
mongod --replSet production-rs --dbpath /data/db --logpath /var/log/mongodb.log --fork --bind_ip 0.0.0.0

// Arbiter server (arbiter.example.com) 
mongod --replSet production-rs --dbpath /data/db --logpath /var/log/mongodb.log --fork --bind_ip 0.0.0.0

// 2. Initialize replica set from primary
rs.initiate({
  _id: "production-rs",
  members: [
    { _id: 0, host: "db1.example.com:27017", priority: 2 },
    { _id: 1, host: "db2.example.com:27017", priority: 1 },
    { _id: 2, host: "db3.example.com:27017", priority: 1 },
    { _id: 3, host: "arbiter.example.com:27017", arbiterOnly: true }
  ]
});

// 3. Verify replica set status
rs.status();

// 4. Monitor replication lag
rs.printSlaveReplicationInfo();

// Production-optimized replica set configuration
const productionConfig = {
  _id: "production-rs",
  version: 1,
  members: [
    {
      _id: 0,
      host: "db-primary-us-east.example.com:27017",
      priority: 3,          // Highest priority
      votes: 1,
      tags: { 
        "datacenter": "us-east", 
        "server_class": "high-performance",
        "backup_target": "primary"
      }
    },
    {
      _id: 1,
      host: "db-secondary-us-east.example.com:27017", 
      priority: 2,          // Secondary priority
      votes: 1,
      tags: { 
        "datacenter": "us-east",
        "server_class": "standard",
        "backup_target": "secondary"
      }
    },
    {
      _id: 2,
      host: "db-secondary-us-west.example.com:27017",
      priority: 1,          // Geographic failover
      votes: 1,
      tags: {
        "datacenter": "us-west",
        "server_class": "standard"
      }
    },
    {
      _id: 3,
      host: "db-hidden-analytics.example.com:27017",
      priority: 0,          // Cannot become primary
      votes: 0,             // Does not vote in elections
      hidden: true,         // Hidden from client routing
      buildIndexes: true,   // Maintains indexes
      tags: {
        "usage": "analytics",
        "datacenter": "us-east"
      }
    }
  ],
  settings: {
    // Election configuration
    electionTimeoutMillis: 10000,      // Time before new election
    heartbeatIntervalMillis: 2000,     // Heartbeat frequency
    catchUpTimeoutMillis: 60000,       // New primary catchup time

    // Write concern settings
    getLastErrorDefaults: {
      w: "majority",                   // Majority write concern default
      j: true,                         // Require journal acknowledgment
      wtimeout: 10000                  // Write timeout
    },

    // Read preference settings
    chainingAllowed: true,             // Allow secondary-to-secondary replication

    // Connection settings
    replicationHeartbeatTimeout: 10000
  }
};

// Apply configuration
rs.reconfig(productionConfig);

// Read preference strategies
class DatabaseConnection {
  constructor() {
    this.client = new MongoClient('mongodb://db1.example.com:27017,db2.example.com:27017,db3.example.com:27017/ecommerce?replicaSet=production-rs');
  }

  // Real-time operations - read from primary for consistency
  async getRealTimeData(collection, query) {
    return await this.client.db()
      .collection(collection)
      .find(query)
      .read(MongoClient.ReadPreference.PRIMARY)
      .toArray();
  }

  // Analytics queries - allow secondary reads for load distribution  
  async getAnalyticsData(collection, pipeline) {
    return await this.client.db()
      .collection(collection)
      .aggregate(pipeline)
      .read(MongoClient.ReadPreference.SECONDARY_PREFERRED)
      .maxTimeMS(300000)  // 5 minute timeout for analytics
      .toArray();
  }

  // Reporting queries - use tagged secondary for dedicated reporting
  async getReportingData(collection, query) {
    return await this.client.db()
      .collection(collection)
      .find(query)
      .read({
        mode: MongoClient.ReadPreference.NEAREST,
        tags: [{ usage: "analytics" }]
      })
      .toArray();
  }

  // Geographically distributed reads
  async getRegionalData(collection, query, region) {
    const readPreference = {
      mode: MongoClient.ReadPreference.NEAREST,
      tags: [{ datacenter: region }],
      maxStalenessMS: 120000  // Allow 2 minutes staleness
    };

    return await this.client.db()
      .collection(collection)
      .find(query)
      .read(readPreference)
      .toArray();
  }
}

-- SQL read replica configuration concepts
-- Primary database for writes and consistent reads
SELECT order_id, status, total_amount
FROM orders@PRIMARY  -- Read from primary for latest data
WHERE customer_id = 12345;

-- Read replicas for analytics and reporting
SELECT 
  DATE(created_at) AS order_date,
  COUNT(*) AS daily_orders,
  SUM(total_amount) AS daily_revenue
FROM orders@SECONDARY_PREFERRED  -- Allow secondary reads
WHERE created_at >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY DATE(created_at);

-- Geographic read routing
SELECT product_id, inventory_count
FROM inventory@NEAREST_DATACENTER('us-west')  -- Route to nearest replica
WHERE product_category = 'electronics';

-- Dedicated analytics server
SELECT customer_id, purchase_behavior_score
FROM customer_analytics@ANALYTICS_REPLICA  -- Dedicated analytics replica
WHERE last_purchase >= CURRENT_DATE - INTERVAL '90 days';

// Replica set failover monitoring
class ReplicaSetMonitor {
  constructor(client) {
    this.client = client;
    this.replicaSetStatus = null;
  }

  async monitorReplicaSetHealth() {
    try {
      // Check replica set status
      const admin = this.client.db().admin();
      this.replicaSetStatus = await admin.command({ replSetGetStatus: 1 });

      return this.analyzeClusterHealth();
    } catch (error) {
      console.error('Failed to get replica set status:', error);
      return { status: 'unknown', error: error.message };
    }
  }

  analyzeClusterHealth() {
    const members = this.replicaSetStatus.members;

    // Count members by state
    const memberStates = {
      primary: members.filter(m => m.state === 1).length,
      secondary: members.filter(m => m.state === 2).length,
      recovering: members.filter(m => m.state === 3).length,
      down: members.filter(m => m.state === 8).length,
      arbiter: members.filter(m => m.state === 7).length
    };

    // Check for healthy primary
    const primaryNode = members.find(m => m.state === 1);

    // Check replication lag
    const maxLag = this.calculateMaxReplicationLag(members);

    // Determine overall cluster health
    let clusterHealth = 'healthy';
    const issues = [];

    if (memberStates.primary === 0) {
      clusterHealth = 'no_primary';
      issues.push('No primary node available');
    } else if (memberStates.primary > 1) {
      clusterHealth = 'split_brain';
      issues.push('Multiple primary nodes detected');
    }

    if (memberStates.down > 0) {
      clusterHealth = 'degraded';
      issues.push(`${memberStates.down} members are down`);
    }

    if (maxLag > 60000) {  // More than 60 seconds lag
      clusterHealth = 'lagged';
      issues.push(`Maximum replication lag: ${maxLag / 1000}s`);
    }

    return {
      status: clusterHealth,
      memberStates: memberStates,
      primary: primaryNode ? primaryNode.name : null,
      maxReplicationLag: maxLag,
      issues: issues,
      timestamp: new Date()
    };
  }

  calculateMaxReplicationLag(members) {
    const primaryNode = members.find(m => m.state === 1);
    if (!primaryNode) return null;

    const primaryOpTime = primaryNode.optimeDate;

    return Math.max(...members
      .filter(m => m.state === 2)  // Secondary nodes only
      .map(member => primaryOpTime - member.optimeDate)
    );
  }
}

// Resilient database operations with retry logic
class ResilientDatabaseClient {
  constructor(connectionString) {
    this.client = new MongoClient(connectionString, {
      replicaSet: 'production-rs',
      maxPoolSize: 50,
      serverSelectionTimeoutMS: 5000,
      socketTimeoutMS: 45000,

      // Retry configuration
      retryWrites: true,
      retryReads: true,

      // Write concern for consistency
      writeConcern: { 
        w: 'majority', 
        j: true, 
        wtimeout: 10000 
      }
    });
  }

  async executeWithRetry(operation, maxRetries = 3) {
    let lastError;

    for (let attempt = 1; attempt <= maxRetries; attempt++) {
      try {
        return await operation();
      } catch (error) {
        lastError = error;

        // Check if error is retryable
        if (this.isRetryableError(error) && attempt < maxRetries) {
          const delay = Math.min(1000 * Math.pow(2, attempt - 1), 5000);
          console.log(`Operation failed (attempt ${attempt}), retrying in ${delay}ms:`, error.message);
          await this.sleep(delay);
          continue;
        }

        throw error;
      }
    }

    throw lastError;
  }

  isRetryableError(error) {
    // Network errors
    if (error.code === 'ECONNRESET' || 
        error.code === 'ENOTFOUND' || 
        error.code === 'ETIMEDOUT') {
      return true;
    }

    // MongoDB specific retryable errors
    const retryableCodes = [
      11600,  // InterruptedAtShutdown
      11602,  // InterruptedDueToReplStateChange  
      10107,  // NotMaster
      13435,  // NotMasterNoSlaveOk
      13436   // NotMasterOrSecondary
    ];

    return retryableCodes.includes(error.code);
  }

  async createOrder(orderData) {
    return await this.executeWithRetry(async () => {
      const session = this.client.startSession();

      try {
        return await session.withTransaction(async () => {
          // Insert order
          const orderResult = await this.client
            .db('ecommerce')
            .collection('orders')
            .insertOne(orderData, { session });

          // Update inventory
          for (const item of orderData.items) {
            await this.client
              .db('ecommerce')
              .collection('inventory')
              .updateOne(
                { 
                  product_id: item.product_id,
                  quantity: { $gte: item.quantity }
                },
                { $inc: { quantity: -item.quantity } },
                { session }
              );
          }

          return orderResult;
        }, {
          readConcern: { level: 'majority' },
          writeConcern: { w: 'majority', j: true }
        });
      } finally {
        await session.endSession();
      }
    });
  }

  sleep(ms) {
    return new Promise(resolve => setTimeout(resolve, ms));
  }
}

-- SQL-style transaction consistency levels
-- Strict consistency - wait for replication to all nodes
INSERT INTO orders (customer_id, total_amount, status)
VALUES (12345, 159.99, 'pending')
WITH CONSISTENCY_LEVEL = 'ALL_REPLICAS',
     TIMEOUT = 10000ms;

-- Majority consistency - wait for majority of nodes
UPDATE inventory 
SET quantity = quantity - 1
WHERE product_id = 'LAPTOP001'
WITH CONSISTENCY_LEVEL = 'MAJORITY',
     JOURNAL_SYNC = true,
     TIMEOUT = 5000ms;

-- Async replication - acknowledge immediately
INSERT INTO user_activity_log (user_id, action, timestamp)
VALUES (12345, 'page_view', NOW())
WITH CONSISTENCY_LEVEL = 'PRIMARY_ONLY',
     ASYNC_REPLICATION = true;

// Write concern strategies for different operations
class TransactionManager {
  constructor(client) {
    this.client = client;
  }

  // Critical financial transactions - strict consistency
  async processPayment(paymentData) {
    const session = this.client.startSession();

    try {
      return await session.withTransaction(async () => {
        // Deduct from account with strict consistency
        await this.client.db('banking')
          .collection('accounts')
          .updateOne(
            { account_id: paymentData.from_account },
            { $inc: { balance: -paymentData.amount } },
            { 
              session,
              writeConcern: { 
                w: "majority",      // Wait for majority
                j: true,            // Wait for journal
                wtimeout: 10000     // 10 second timeout
              }
            }
          );

        // Credit target account
        await this.client.db('banking')
          .collection('accounts')
          .updateOne(
            { account_id: paymentData.to_account },
            { $inc: { balance: paymentData.amount } },
            { session, writeConcern: { w: "majority", j: true, wtimeout: 10000 } }
          );

        // Log transaction
        await this.client.db('banking')
          .collection('transaction_log')
          .insertOne({
            from_account: paymentData.from_account,
            to_account: paymentData.to_account,
            amount: paymentData.amount,
            timestamp: new Date(),
            status: 'completed'
          }, { 
            session, 
            writeConcern: { w: "majority", j: true, wtimeout: 10000 }
          });

      }, {
        readConcern: { level: 'majority' },
        writeConcern: { w: 'majority', j: true, wtimeout: 15000 }
      });
    } finally {
      await session.endSession();
    }
  }

  // Standard business operations - balanced consistency
  async createOrder(orderData) {
    return await this.client.db('ecommerce')
      .collection('orders')
      .insertOne(orderData, {
        writeConcern: { 
          w: "majority",    // Majority of voting members
          j: true,          // Journal acknowledgment  
          wtimeout: 5000    // 5 second timeout
        }
      });
  }

  // Analytics and logging - performance optimized
  async logUserActivity(activityData) {
    return await this.client.db('analytics')
      .collection('user_events')
      .insertOne(activityData, {
        writeConcern: { 
          w: 1,           // Primary only
          j: false,       // No journal wait
          wtimeout: 1000  // Quick timeout
        }
      });
  }

  // Bulk operations - optimized for throughput
  async bulkInsertAnalytics(documents) {
    return await this.client.db('analytics')
      .collection('events')
      .insertMany(documents, {
        ordered: false,   // Allow parallel inserts
        writeConcern: { 
          w: 1,          // Primary acknowledgment only
          j: false       // No journal synchronization
        }
      });
  }
}

// Automated backup system
class ReplicaSetBackupManager {
  constructor(client, config) {
    this.client = client;
    this.config = config;
  }

  async performIncrementalBackup() {
    // Use oplog for incremental backups
    const admin = this.client.db().admin();
    const oplogCollection = this.client.db('local').collection('oplog.rs');

    // Get last backup timestamp
    const lastBackupTime = await this.getLastBackupTimestamp();

    // Query oplog entries since last backup
    const oplogEntries = await oplogCollection.find({
      ts: { $gt: lastBackupTime },
      ns: { $not: /^(admin\.|config\.)/ }  // Skip system databases
    }).toArray();

    // Process and store oplog entries
    await this.storeIncrementalBackup(oplogEntries);

    // Update backup timestamp
    await this.updateLastBackupTimestamp();

    return {
      entriesProcessed: oplogEntries.length,
      backupTime: new Date(),
      type: 'incremental'
    };
  }

  async performFullBackup() {
    const databases = await this.client.db().admin().listDatabases();
    const backupResults = [];

    for (const dbInfo of databases.databases) {
      if (this.shouldBackupDatabase(dbInfo.name)) {
        const result = await this.backupDatabase(dbInfo.name);
        backupResults.push(result);
      }
    }

    return {
      databases: backupResults,
      backupTime: new Date(),
      type: 'full'
    };
  }

  async backupDatabase(databaseName) {
    const db = this.client.db(databaseName);
    const collections = await db.listCollections().toArray();
    const collectionBackups = [];

    for (const collInfo of collections) {
      if (collInfo.type === 'collection') {
        const documents = await db.collection(collInfo.name).find({}).toArray();
        const indexes = await db.collection(collInfo.name).listIndexes().toArray();

        await this.storeCollectionBackup(databaseName, collInfo.name, {
          documents: documents,
          indexes: indexes,
          options: collInfo.options
        });

        collectionBackups.push({
          name: collInfo.name,
          documentCount: documents.length,
          indexCount: indexes.length
        });
      }
    }

    return {
      database: databaseName,
      collections: collectionBackups
    };
  }

  shouldBackupDatabase(dbName) {
    const systemDatabases = ['admin', 'config', 'local'];
    return !systemDatabases.includes(dbName);
  }
}

-- SQL backup strategies equivalent
-- Full database backup
BACKUP DATABASE ecommerce 
TO '/backups/ecommerce_full_2025-08-28.bak'
WITH FORMAT, 
     INIT,
     COMPRESSION,
     CHECKSUM;

-- Transaction log backup for point-in-time recovery
BACKUP LOG ecommerce
TO '/backups/ecommerce_log_2025-08-28_10-15.trn'
WITH COMPRESSION;

-- Differential backup
BACKUP DATABASE ecommerce
TO '/backups/ecommerce_diff_2025-08-28.bak'
WITH DIFFERENTIAL,
     COMPRESSION,
     CHECKSUM;

-- Point-in-time restore
RESTORE DATABASE ecommerce_recovery
FROM '/backups/ecommerce_full_2025-08-28.bak'
WITH NORECOVERY,
     MOVE 'ecommerce' TO '/data/ecommerce_recovery.mdf';

RESTORE LOG ecommerce_recovery  
FROM '/backups/ecommerce_log_2025-08-28_10-15.trn'
WITH RECOVERY,
     STOPAT = '2025-08-28 10:14:30.000';

// Comprehensive replica set monitoring
class ReplicaSetPerformanceMonitor {
  constructor(client) {
    this.client = client;
    this.metrics = new Map();
  }

  async collectMetrics() {
    const metrics = {
      replicationLag: await this.measureReplicationLag(),
      oplogStats: await this.getOplogStatistics(), 
      connectionStats: await this.getConnectionStatistics(),
      memberHealth: await this.assessMemberHealth(),
      throughputStats: await this.measureThroughput()
    };

    this.metrics.set(Date.now(), metrics);
    return metrics;
  }

  async measureReplicationLag() {
    const replSetStatus = await this.client.db().admin().command({ replSetGetStatus: 1 });
    const primary = replSetStatus.members.find(m => m.state === 1);

    if (!primary) return null;

    const secondaries = replSetStatus.members.filter(m => m.state === 2);
    const lagStats = secondaries.map(secondary => ({
      member: secondary.name,
      lag: primary.optimeDate - secondary.optimeDate,
      state: secondary.stateStr,
      health: secondary.health
    }));

    return {
      maxLag: Math.max(...lagStats.map(s => s.lag)),
      avgLag: lagStats.reduce((sum, s) => sum + s.lag, 0) / lagStats.length,
      members: lagStats
    };
  }

  async getOplogStatistics() {
    const oplogStats = await this.client.db('local').collection('oplog.rs').stats();
    const firstEntry = await this.client.db('local').collection('oplog.rs')
      .findOne({}, { sort: { ts: 1 } });
    const lastEntry = await this.client.db('local').collection('oplog.rs')  
      .findOne({}, { sort: { ts: -1 } });

    if (!firstEntry || !lastEntry) return null;

    const oplogSpan = lastEntry.ts.getHighBits() - firstEntry.ts.getHighBits();

    return {
      size: oplogStats.size,
      count: oplogStats.count,
      avgObjSize: oplogStats.avgObjSize,
      oplogSpanHours: oplogSpan / 3600,
      utilizationPercent: (oplogStats.size / oplogStats.maxSize) * 100
    };
  }

  async measureThroughput() {
    const serverStatus = await this.client.db().admin().command({ serverStatus: 1 });

    return {
      insertRate: serverStatus.metrics?.document?.inserted || 0,
      updateRate: serverStatus.metrics?.document?.updated || 0, 
      deleteRate: serverStatus.metrics?.document?.deleted || 0,
      queryRate: serverStatus.metrics?.queryExecutor?.scanned || 0,
      connectionCount: serverStatus.connections?.current || 0
    };
  }

  generateHealthReport() {
    const latestMetrics = Array.from(this.metrics.values()).pop();
    if (!latestMetrics) return null;

    const healthScore = this.calculateHealthScore(latestMetrics);
    const recommendations = this.generateRecommendations(latestMetrics);

    return {
      overall_health: healthScore > 80 ? 'excellent' : 
                     healthScore > 60 ? 'good' : 
                     healthScore > 40 ? 'fair' : 'poor',
      health_score: healthScore,
      metrics: latestMetrics,
      recommendations: recommendations,
      timestamp: new Date()
    };
  }

  calculateHealthScore(metrics) {
    let score = 100;

    // Penalize high replication lag
    if (metrics.replicationLag?.maxLag > 60000) {
      score -= 30; // > 60 seconds lag
    } else if (metrics.replicationLag?.maxLag > 10000) {
      score -= 15; // > 10 seconds lag
    }

    // Penalize unhealthy members
    const unhealthyMembers = metrics.memberHealth?.filter(m => m.health !== 1).length || 0;
    score -= unhealthyMembers * 20;

    // Penalize high oplog utilization
    if (metrics.oplogStats?.utilizationPercent > 80) {
      score -= 15;
    }

    return Math.max(0, score);
  }
}

-- QueryLeaf automatically handles replica set operations
-- Connection configuration handles failover transparently
CONNECT TO mongodb_cluster WITH (
  hosts = 'db1.example.com:27017,db2.example.com:27017,db3.example.com:27017',
  replica_set = 'production-rs',
  read_preference = 'primaryPreferred',
  write_concern = 'majority'
);

-- Read operations automatically route based on preferences
SELECT 
  order_id,
  customer_id, 
  total_amount,
  status,
  created_at
FROM orders 
WHERE status = 'pending'
  AND created_at >= CURRENT_DATE - INTERVAL '1 day'
READ_PREFERENCE = 'secondary';  -- QueryLeaf extension for read routing

-- Write operations use configured write concerns
INSERT INTO orders (
  customer_id,
  items,
  total_amount,
  status
) VALUES (
  OBJECTID('64f1a2c4567890abcdef5678'),
  '[{"product": "laptop", "quantity": 1, "price": 1200}]'::jsonb,
  1200.00,
  'pending'
)
WITH WRITE_CONCERN = '{ w: "majority", j: true, wtimeout: 10000 }';

-- Analytics queries can target specific replica members
SELECT 
  DATE_TRUNC('hour', created_at) AS hour,
  COUNT(*) AS order_count,
  SUM(total_amount) AS revenue
FROM orders 
WHERE created_at >= CURRENT_DATE - INTERVAL '7 days'
GROUP BY DATE_TRUNC('hour', created_at)
READ_PREFERENCE = 'nearest', TAGS = '{ "usage": "analytics" }';

-- QueryLeaf provides:
-- 1. Automatic failover handling in SQL connections
-- 2. Transparent read preference management  
-- 3. Write concern configuration through SQL
-- 4. Connection pooling optimized for replica sets
-- 5. Monitoring integration for replica set health