specweave 0.30.12 → 0.30.14

package/plugins/specweave-kafka/benchmarks/kafka-throughput.benchmark.ts

@@ -0,0 +1,551 @@
+/**
+ * Kafka Performance Benchmarks
+ *
+ * Measures throughput and latency for:
+ * - Producer performance (msgs/sec, MB/sec)
+ * - Consumer performance (msgs/sec, lag)
+ * - End-to-end latency (p50, p95, p99)
+ * - Batch processing efficiency
+ * - Compression impact
+ * - Concurrent operations
+ *
+ * Target: 100K+ msgs/sec throughput
+ *
+ * @benchmark
+ */
+
+import { Kafka, Producer, Consumer, CompressionTypes } from 'kafkajs';
+import { performance } from 'perf_hooks';
+import { v4 as uuidv4 } from 'uuid';
+import * as fs from 'fs';
+import * as path from 'path';
+
+interface BenchmarkResult {
+  name: string;
+  duration: number;
+  messageCount: number;
+  throughput: number; // msgs/sec
+  bytesPerSecond: number;
+  latencyP50?: number;
+  latencyP95?: number;
+  latencyP99?: number;
+  avgLatency?: number;
+}
+
+class KafkaBenchmark {
+  private kafka: Kafka;
+  private results: BenchmarkResult[] = [];
+
+  constructor() {
+    this.kafka = new Kafka({
+      clientId: 'benchmark-client',
+      brokers: process.env.KAFKA_BROKERS?.split(',') || ['localhost:9092'],
+      retry: {
+        retries: 3,
+      },
+    });
+  }
+
+  async runAll(): Promise<void> {
+    console.log('🚀 Starting Kafka Performance Benchmarks...\n');
+
+    await this.benchmarkProducerThroughput();
+    await this.benchmarkConsumerThroughput();
+    await this.benchmarkEndToEndLatency();
+    await this.benchmarkBatchSizes();
+    await this.benchmarkCompression();
+    await this.benchmarkConcurrentProducers();
+
+    this.generateReport();
+  }
+
+  /**
+   * Benchmark 1: Producer Throughput
+   * Target: 100K+ msgs/sec
+   */
+  private async benchmarkProducerThroughput(): Promise<void> {
+    console.log('📊 Benchmark 1: Producer Throughput\n');
+
+    const topic = `bench-producer-${uuidv4()}`;
+    const messageCount = 100000;
+    const messageSize = 1024; // 1KB
+
+    const admin = this.kafka.admin();
+    await admin.connect();
+    await admin.createTopics({
+      topics: [{ topic, numPartitions: 10 }],
+    });
+    await admin.disconnect();
+
+    const producer = this.kafka.producer({
+      allowAutoTopicCreation: false,
+      idempotent: true,
+    });
+
+    await producer.connect();
+
+    const message = {
+      value: Buffer.alloc(messageSize).toString('base64'),
+    };
+
+    const startTime = performance.now();
+
+    // Send in batches for better throughput
+    const batchSize = 1000;
+    for (let i = 0; i < messageCount; i += batchSize) {
+      const batch = Array(Math.min(batchSize, messageCount - i)).fill(message);
+      await producer.send({
+        topic,
+        messages: batch,
+      });
+
+      if ((i + batchSize) % 10000 === 0) {
+        process.stdout.write(`\rProgress: ${i + batchSize}/${messageCount}`);
+      }
+    }
+
+    const endTime = performance.now();
+    const duration = (endTime - startTime) / 1000; // seconds
+
+    await producer.disconnect();
+
+    const throughput = messageCount / duration;
+    const bytesPerSecond = (messageCount * messageSize) / duration;
+
+    this.results.push({
+      name: 'Producer Throughput',
+      duration,
+      messageCount,
+      throughput,
+      bytesPerSecond,
+    });
+
+    console.log(`\n✅ Completed: ${throughput.toFixed(0)} msgs/sec, ${(bytesPerSecond / 1024 / 1024).toFixed(2)} MB/sec\n`);
+  }
+
+  /**
+   * Benchmark 2: Consumer Throughput
+   */
+  private async benchmarkConsumerThroughput(): Promise<void> {
+    console.log('📊 Benchmark 2: Consumer Throughput\n');
+
+    const topic = `bench-consumer-${uuidv4()}`;
+    const messageCount = 50000;
+    const messageSize = 1024;
+
+    const admin = this.kafka.admin();
+    await admin.connect();
+    await admin.createTopics({
+      topics: [{ topic, numPartitions: 10 }],
+    });
+    await admin.disconnect();
+
+    // Produce messages first
+    const producer = this.kafka.producer();
+    await producer.connect();
+
+    const message = {
+      value: Buffer.alloc(messageSize).toString('base64'),
+    };
+
+    for (let i = 0; i < messageCount; i += 1000) {
+      const batch = Array(1000).fill(message);
+      await producer.send({ topic, messages: batch });
+    }
+
+    await producer.disconnect();
+
+    // Benchmark consumption
+    const consumer = this.kafka.consumer({
+      groupId: `bench-group-${uuidv4()}`,
+    });
+
+    await consumer.connect();
+    await consumer.subscribe({ topic, fromBeginning: true });
+
+    let consumedCount = 0;
+    const startTime = performance.now();
+
+    await new Promise<void>((resolve) => {
+      consumer.run({
+        eachBatch: async ({ batch }) => {
+          consumedCount += batch.messages.length;
+
+          if (consumedCount >= messageCount) {
+            resolve();
+          }
+        },
+      });
+    });
+
+    const endTime = performance.now();
+    const duration = (endTime - startTime) / 1000;
+
+    await consumer.disconnect();
+
+    const throughput = consumedCount / duration;
+    const bytesPerSecond = (consumedCount * messageSize) / duration;
+
+    this.results.push({
+      name: 'Consumer Throughput',
+      duration,
+      messageCount: consumedCount,
+      throughput,
+      bytesPerSecond,
+    });
+
+    console.log(`✅ Completed: ${throughput.toFixed(0)} msgs/sec, ${(bytesPerSecond / 1024 / 1024).toFixed(2)} MB/sec\n`);
+  }
+
+  /**
+   * Benchmark 3: End-to-End Latency
+   */
+  private async benchmarkEndToEndLatency(): Promise<void> {
+    console.log('📊 Benchmark 3: End-to-End Latency\n');
+
+    const topic = `bench-latency-${uuidv4()}`;
+    const messageCount = 10000;
+
+    const admin = this.kafka.admin();
+    await admin.connect();
+    await admin.createTopics({
+      topics: [{ topic, numPartitions: 1 }],
+    });
+    await admin.disconnect();
+
+    const producer = this.kafka.producer();
+    const consumer = this.kafka.consumer({
+      groupId: `bench-latency-group-${uuidv4()}`,
+    });
+
+    await producer.connect();
+    await consumer.connect();
+    await consumer.subscribe({ topic, fromBeginning: true });
+
+    const latencies: number[] = [];
+    const timestamps = new Map<string, number>();
+
+    const consumePromise = new Promise<void>((resolve) => {
+      consumer.run({
+        eachMessage: async ({ message }) => {
+          const id = message.value!.toString();
+          const sendTime = timestamps.get(id);
+
+          if (sendTime) {
+            const latency = performance.now() - sendTime;
+            latencies.push(latency);
+
+            if (latencies.length >= messageCount) {
+              resolve();
+            }
+          }
+        },
+      });
+    });
+
+    // Send messages with timestamps
+    for (let i = 0; i < messageCount; i++) {
+      const id = `msg-${i}`;
+      timestamps.set(id, performance.now());
+
+      await producer.send({
+        topic,
+        messages: [{ value: id }],
+      });
+
+      if (i % 1000 === 0) {
+        process.stdout.write(`\rProgress: ${i}/${messageCount}`);
+      }
+    }
+
+    await consumePromise;
+
+    await producer.disconnect();
+    await consumer.disconnect();
+
+    // Calculate percentiles
+    latencies.sort((a, b) => a - b);
+
+    const p50 = latencies[Math.floor(latencies.length * 0.50)];
+    const p95 = latencies[Math.floor(latencies.length * 0.95)];
+    const p99 = latencies[Math.floor(latencies.length * 0.99)];
+    const avg = latencies.reduce((sum, l) => sum + l, 0) / latencies.length;
+
+    this.results.push({
+      name: 'End-to-End Latency',
+      duration: 0,
+      messageCount,
+      throughput: 0,
+      bytesPerSecond: 0,
+      latencyP50: p50,
+      latencyP95: p95,
+      latencyP99: p99,
+      avgLatency: avg,
+    });
+
+    console.log(`\n✅ Latency - p50: ${p50.toFixed(2)}ms, p95: ${p95.toFixed(2)}ms, p99: ${p99.toFixed(2)}ms\n`);
+  }
+
+  /**
+   * Benchmark 4: Batch Size Impact
+   */
+  private async benchmarkBatchSizes(): Promise<void> {
+    console.log('📊 Benchmark 4: Batch Size Impact\n');
+
+    const batchSizes = [10, 100, 500, 1000, 5000];
+
+    for (const batchSize of batchSizes) {
+      const topic = `bench-batch-${batchSize}-${uuidv4()}`;
+      const messageCount = 50000;
+
+      const admin = this.kafka.admin();
+      await admin.connect();
+      await admin.createTopics({
+        topics: [{ topic, numPartitions: 5 }],
+      });
+      await admin.disconnect();
+
+      const producer = this.kafka.producer();
+      await producer.connect();
+
+      const startTime = performance.now();
+
+      for (let i = 0; i < messageCount; i += batchSize) {
+        const batch = Array(Math.min(batchSize, messageCount - i))
+          .fill({ value: 'benchmark-data' });
+
+        await producer.send({ topic, messages: batch });
+      }
+
+      const endTime = performance.now();
+      const duration = (endTime - startTime) / 1000;
+
+      await producer.disconnect();
+
+      const throughput = messageCount / duration;
+
+      this.results.push({
+        name: `Batch Size ${batchSize}`,
+        duration,
+        messageCount,
+        throughput,
+        bytesPerSecond: 0,
+      });
+
+      console.log(`  Batch ${batchSize}: ${throughput.toFixed(0)} msgs/sec`);
+    }
+
+    console.log();
+  }
+
+  /**
+   * Benchmark 5: Compression Impact
+   */
+  private async benchmarkCompression(): Promise<void> {
+    console.log('📊 Benchmark 5: Compression Impact\n');
+
+    const compressionTypes = [
+      { type: CompressionTypes.None, name: 'None' },
+      { type: CompressionTypes.GZIP, name: 'GZIP' },
+      { type: CompressionTypes.Snappy, name: 'Snappy' },
+      { type: CompressionTypes.LZ4, name: 'LZ4' },
+      { type: CompressionTypes.ZSTD, name: 'ZSTD' },
+    ];
+
+    for (const compression of compressionTypes) {
+      const topic = `bench-compression-${compression.name}-${uuidv4()}`;
+      const messageCount = 10000;
+      const messageSize = 10240; // 10KB (compressible)
+
+      const admin = this.kafka.admin();
+      await admin.connect();
+      await admin.createTopics({
+        topics: [{ topic, numPartitions: 3 }],
+      });
+      await admin.disconnect();
+
+      const producer = this.kafka.producer();
+      await producer.connect();
+
+      const message = {
+        value: 'A'.repeat(messageSize), // Highly compressible
+      };
+
+      const startTime = performance.now();
+
+      for (let i = 0; i < messageCount; i += 100) {
+        const batch = Array(100).fill(message);
+        await producer.send({
+          topic,
+          compression: compression.type,
+          messages: batch,
+        });
+      }
+
+      const endTime = performance.now();
+      const duration = (endTime - startTime) / 1000;
+
+      await producer.disconnect();
+
+      const throughput = messageCount / duration;
+      const bytesPerSecond = (messageCount * messageSize) / duration;
+
+      this.results.push({
+        name: `Compression: ${compression.name}`,
+        duration,
+        messageCount,
+        throughput,
+        bytesPerSecond,
+      });
+
+      console.log(`  ${compression.name}: ${throughput.toFixed(0)} msgs/sec, ${(bytesPerSecond / 1024 / 1024).toFixed(2)} MB/sec`);
+    }
+
+    console.log();
+  }
+
+  /**
+   * Benchmark 6: Concurrent Producers
+   */
+  private async benchmarkConcurrentProducers(): Promise<void> {
+    console.log('📊 Benchmark 6: Concurrent Producers\n');
+
+    const topic = `bench-concurrent-${uuidv4()}`;
+    const producerCount = 10;
+    const messagesPerProducer = 10000;
+
+    const admin = this.kafka.admin();
+    await admin.connect();
+    await admin.createTopics({
+      topics: [{ topic, numPartitions: 10 }],
+    });
+    await admin.disconnect();
+
+    const startTime = performance.now();
+
+    const producers = await Promise.all(
+      Array.from({ length: producerCount }, async () => {
+        const producer = this.kafka.producer();
+        await producer.connect();
+        return producer;
+      })
+    );
+
+    await Promise.all(
+      producers.map(async (producer) => {
+        for (let i = 0; i < messagesPerProducer; i += 100) {
+          const batch = Array(100).fill({ value: 'concurrent-test' });
+          await producer.send({ topic, messages: batch });
+        }
+      })
+    );
+
+    const endTime = performance.now();
+    const duration = (endTime - startTime) / 1000;
+
+    await Promise.all(producers.map(p => p.disconnect()));
+
+    const totalMessages = producerCount * messagesPerProducer;
+    const throughput = totalMessages / duration;
+
+    this.results.push({
+      name: `Concurrent Producers (${producerCount})`,
+      duration,
+      messageCount: totalMessages,
+      throughput,
+      bytesPerSecond: 0,
+    });
+
+    console.log(`✅ Completed: ${throughput.toFixed(0)} msgs/sec with ${producerCount} concurrent producers\n`);
+  }
+
+  /**
+   * Generate benchmark report
+   */
+  private generateReport(): void {
+    console.log('\n' + '='.repeat(80));
+    console.log('📈 BENCHMARK RESULTS SUMMARY');
+    console.log('='.repeat(80) + '\n');
+
+    this.results.forEach((result) => {
+      console.log(`${result.name}:`);
+      console.log(`  Messages: ${result.messageCount.toLocaleString()}`);
+
+      if (result.throughput > 0) {
+        console.log(`  Throughput: ${result.throughput.toFixed(0)} msgs/sec`);
+      }
+
+      if (result.bytesPerSecond > 0) {
+        console.log(`  Bandwidth: ${(result.bytesPerSecond / 1024 / 1024).toFixed(2)} MB/sec`);
+      }
+
+      if (result.latencyP50) {
+        console.log(`  Latency p50: ${result.latencyP50.toFixed(2)}ms`);
+        console.log(`  Latency p95: ${result.latencyP95!.toFixed(2)}ms`);
+        console.log(`  Latency p99: ${result.latencyP99!.toFixed(2)}ms`);
+      }
+
+      if (result.duration > 0) {
+        console.log(`  Duration: ${result.duration.toFixed(2)}s`);
+      }
+
+      console.log();
+    });
+
+    // Save to file
+    const reportDir = path.join(process.cwd(), 'benchmark-results');
+    if (!fs.existsSync(reportDir)) {
+      fs.mkdirSync(reportDir, { recursive: true });
+    }
+
+    const timestamp = new Date().toISOString().replace(/:/g, '-');
+    const reportPath = path.join(reportDir, `benchmark-${timestamp}.json`);
+
+    fs.writeFileSync(reportPath, JSON.stringify({
+      timestamp: new Date().toISOString(),
+      results: this.results,
+      environment: {
+        nodeVersion: process.version,
+        platform: process.platform,
+        kafkaBrokers: process.env.KAFKA_BROKERS || 'localhost:9092',
+      },
+    }, null, 2));
+
+    console.log(`📄 Full report saved to: ${reportPath}\n`);
+
+    // Performance validation
+    const producerThroughput = this.results.find(r => r.name === 'Producer Throughput');
+    if (producerThroughput && producerThroughput.throughput >= 100000) {
+      console.log('✅ PASS: Producer throughput meets 100K+ msgs/sec target\n');
+    } else {
+      console.log('⚠️  WARNING: Producer throughput below 100K msgs/sec target\n');
+    }
+  }
+}
+
+// Run benchmarks
+(async () => {
+  const benchmark = new KafkaBenchmark();
+  await benchmark.runAll();
+  process.exit(0);
+})();
+
+/**
+ * Benchmark Summary:
+ *
+ * 1. Producer Throughput - 100K+ msgs/sec target
+ * 2. Consumer Throughput - Maximum consumption rate
+ * 3. End-to-End Latency - p50, p95, p99 percentiles
+ * 4. Batch Size Impact - Optimal batch sizing
+ * 5. Compression Impact - Codec performance comparison
+ * 6. Concurrent Producers - Scalability validation
+ *
+ * Expected Results:
+ * - Producer: 100K-500K msgs/sec
+ * - Consumer: 100K-300K msgs/sec
+ * - Latency p50: <10ms
+ * - Latency p95: <50ms
+ * - Latency p99: <100ms
+ *
+ * Run: npm run benchmark
+ */