specweave 0.18.0 → 0.20.0

package/plugins/specweave-kafka/lib/multi-cluster/health-aggregator.ts

@@ -0,0 +1,380 @@
+/**
+ * Health Aggregator
+ *
+ * Aggregates health metrics across multiple Kafka clusters
+ *
+ * @module health-aggregator
+ */
+
+import { Admin } from 'kafkajs';
+import { ClusterSwitcher } from './cluster-switcher';
+import { ClusterConfig } from './cluster-config-manager';
+
+/**
+ * Cluster Health Metrics
+ */
+export interface ClusterHealthMetrics {
+  /** Cluster ID */
+  clusterId: string;
+  /** Cluster name */
+  clusterName: string;
+  /** Environment */
+  environment: string;
+  /** Overall health status */
+  status: 'healthy' | 'degraded' | 'down';
+  /** Number of brokers */
+  brokerCount: number;
+  /** Number of online brokers */
+  onlineBrokerCount: number;
+  /** Number of topics */
+  topicCount: number;
+  /** Number of partitions */
+  partitionCount: number;
+  /** Under-replicated partitions */
+  underReplicatedPartitions: number;
+  /** Offline partitions */
+  offlinePartitions: number;
+  /** Consumer groups */
+  consumerGroupCount: number;
+  /** Controller ID */
+  controllerId?: number;
+  /** Last check timestamp */
+  lastCheck: Date;
+  /** Error message (if down) */
+  error?: string;
+}
+
+/**
+ * Aggregated Health Summary
+ */
+export interface AggregatedHealthSummary {
+  /** Total clusters */
+  totalClusters: number;
+  /** Healthy clusters */
+  healthyClusters: number;
+  /** Degraded clusters */
+  degradedClusters: number;
+  /** Down clusters */
+  downClusters: number;
+  /** Total brokers across all clusters */
+  totalBrokers: number;
+  /** Total topics across all clusters */
+  totalTopics: number;
+  /** Total partitions across all clusters */
+  totalPartitions: number;
+  /** Total under-replicated partitions across all clusters */
+  totalUnderReplicatedPartitions: number;
+  /** Per-cluster metrics */
+  clusterMetrics: ClusterHealthMetrics[];
+  /** Last update timestamp */
+  lastUpdate: Date;
+}
+
+/**
+ * Health Aggregator
+ *
+ * Collects and aggregates health metrics from multiple Kafka clusters
+ */
+export class HealthAggregator {
+  private switcher: ClusterSwitcher;
+
+  constructor(switcher: ClusterSwitcher) {
+    this.switcher = switcher;
+  }
+
+  /**
+   * Collect health metrics for a single cluster
+   */
+  async collectClusterHealth(clusterId: string): Promise<ClusterHealthMetrics> {
+    try {
+      // Get cluster config
+      const clusters = this.switcher.listClusters();
+      const clusterConfig = clusters.find((c) => c.id === clusterId);
+      if (!clusterConfig) {
+        throw new Error(`Cluster "${clusterId}" not found`);
+      }
+
+      // Execute health check on this cluster
+      const metrics = await this.switcher.executeOn(clusterId, async (kafka) => {
+        const admin = kafka.admin();
+        await admin.connect();
+
+        try {
+          // Get cluster metadata
+          const cluster = await admin.describeCluster();
+          const topics = await admin.listTopics();
+          const groups = await admin.listGroups();
+
+          // Get topic metadata (for partition counts)
+          const metadata = await admin.fetchTopicMetadata({ topics });
+
+          // Calculate partition counts
+          let totalPartitions = 0;
+          let underReplicatedPartitions = 0;
+          let offlinePartitions = 0;
+
+          for (const topic of metadata.topics) {
+            for (const partition of topic.partitions) {
+              totalPartitions++;
+
+              // Under-replicated: ISR < replicas
+              if (partition.isr.length < partition.replicas.length) {
+                underReplicatedPartitions++;
+              }
+
+              // Offline: ISR empty
+              if (partition.isr.length === 0) {
+                offlinePartitions++;
+              }
+            }
+          }
+
+          return {
+            clusterId,
+            clusterName: clusterConfig.name,
+            environment: clusterConfig.environment,
+            status: this.determineStatus(underReplicatedPartitions, offlinePartitions),
+            brokerCount: cluster.brokers.length,
+            onlineBrokerCount: cluster.brokers.length, // All in describeCluster are online
+            topicCount: topics.length,
+            partitionCount: totalPartitions,
+            underReplicatedPartitions,
+            offlinePartitions,
+            consumerGroupCount: groups.groups.length,
+            controllerId: cluster.controller ? Number(cluster.controller) : undefined,
+            lastCheck: new Date(),
+          };
+        } finally {
+          await admin.disconnect();
+        }
+      });
+
+      return metrics as ClusterHealthMetrics;
+    } catch (error) {
+      // Return minimal metrics with down status
+      const clusters = this.switcher.listClusters();
+      const clusterConfig = clusters.find((c) => c.id === clusterId)!;
+
+      return {
+        clusterId,
+        clusterName: clusterConfig.name,
+        environment: clusterConfig.environment,
+        status: 'down',
+        brokerCount: 0,
+        onlineBrokerCount: 0,
+        topicCount: 0,
+        partitionCount: 0,
+        underReplicatedPartitions: 0,
+        offlinePartitions: 0,
+        consumerGroupCount: 0,
+        lastCheck: new Date(),
+        error: (error as Error).message,
+      };
+    }
+  }
+
+  /**
+   * Determine cluster status based on metrics
+   */
+  private determineStatus(
+    underReplicatedPartitions: number,
+    offlinePartitions: number
+  ): 'healthy' | 'degraded' | 'down' {
+    if (offlinePartitions > 0) {
+      return 'down'; // Any offline partitions = cluster down
+    }
+
+    if (underReplicatedPartitions > 0) {
+      return 'degraded'; // Under-replicated = degraded
+    }
+
+    return 'healthy';
+  }
+
+  /**
+   * Collect health metrics for all clusters
+   */
+  async collectAllClusters(): Promise<ClusterHealthMetrics[]> {
+    const clusters = this.switcher.listClusters();
+    const promises = clusters.map((c) => this.collectClusterHealth(c.id));
+
+    return Promise.all(promises);
+  }
+
+  /**
+   * Aggregate health across all clusters
+   */
+  async aggregateHealth(): Promise<AggregatedHealthSummary> {
+    const clusterMetrics = await this.collectAllClusters();
+
+    // Calculate totals
+    let totalBrokers = 0;
+    let totalTopics = 0;
+    let totalPartitions = 0;
+    let totalUnderReplicatedPartitions = 0;
+    let healthyClusters = 0;
+    let degradedClusters = 0;
+    let downClusters = 0;
+
+    for (const metrics of clusterMetrics) {
+      totalBrokers += metrics.brokerCount;
+      totalTopics += metrics.topicCount;
+      totalPartitions += metrics.partitionCount;
+      totalUnderReplicatedPartitions += metrics.underReplicatedPartitions;
+
+      if (metrics.status === 'healthy') healthyClusters++;
+      else if (metrics.status === 'degraded') degradedClusters++;
+      else if (metrics.status === 'down') downClusters++;
+    }
+
+    return {
+      totalClusters: clusterMetrics.length,
+      healthyClusters,
+      degradedClusters,
+      downClusters,
+      totalBrokers,
+      totalTopics,
+      totalPartitions,
+      totalUnderReplicatedPartitions,
+      clusterMetrics,
+      lastUpdate: new Date(),
+    };
+  }
+
+  /**
+   * Get health summary as formatted text
+   */
+  async getHealthSummaryText(): Promise<string> {
+    const summary = await this.aggregateHealth();
+
+    const lines: string[] = [];
+    lines.push('=== Kafka Multi-Cluster Health Summary ===');
+    lines.push('');
+    lines.push(`Total Clusters: ${summary.totalClusters}`);
+    lines.push(`  ✓ Healthy:  ${summary.healthyClusters}`);
+    lines.push(`  ⚠ Degraded: ${summary.degradedClusters}`);
+    lines.push(`  ✗ Down:     ${summary.downClusters}`);
+    lines.push('');
+    lines.push(`Total Brokers:  ${summary.totalBrokers}`);
+    lines.push(`Total Topics:   ${summary.totalTopics}`);
+    lines.push(`Total Partitions: ${summary.totalPartitions}`);
+    if (summary.totalUnderReplicatedPartitions > 0) {
+      lines.push(`⚠ Under-Replicated Partitions: ${summary.totalUnderReplicatedPartitions}`);
+    }
+    lines.push('');
+    lines.push('=== Per-Cluster Status ===');
+
+    for (const cluster of summary.clusterMetrics) {
+      const icon =
+        cluster.status === 'healthy' ? '✓' : cluster.status === 'degraded' ? '⚠' : '✗';
+
+      lines.push('');
+      lines.push(
+        `${icon} ${cluster.clusterName} (${cluster.environment}) - ${cluster.status.toUpperCase()}`
+      );
+      lines.push(`  Brokers: ${cluster.onlineBrokerCount}/${cluster.brokerCount}`);
+      lines.push(`  Topics: ${cluster.topicCount}`);
+      lines.push(`  Partitions: ${cluster.partitionCount}`);
+
+      if (cluster.underReplicatedPartitions > 0) {
+        lines.push(`  ⚠ Under-Replicated: ${cluster.underReplicatedPartitions}`);
+      }
+
+      if (cluster.offlinePartitions > 0) {
+        lines.push(`  ✗ Offline: ${cluster.offlinePartitions}`);
+      }
+
+      if (cluster.error) {
+        lines.push(`  Error: ${cluster.error}`);
+      }
+    }
+
+    lines.push('');
+    lines.push(`Last Updated: ${summary.lastUpdate.toISOString()}`);
+
+    return lines.join('\n');
+  }
+
+  /**
+   * Check if any cluster is unhealthy
+   */
+  async hasUnhealthyClusters(): Promise<boolean> {
+    const summary = await this.aggregateHealth();
+    return summary.degradedClusters > 0 || summary.downClusters > 0;
+  }
+
+  /**
+   * Get list of unhealthy clusters
+   */
+  async getUnhealthyClusters(): Promise<ClusterHealthMetrics[]> {
+    const clusterMetrics = await this.collectAllClusters();
+    return clusterMetrics.filter((c) => c.status !== 'healthy');
+  }
+}
+
+/**
+ * Example Usage: Health Check All Clusters
+ *
+ * ```typescript
+ * const switcher = new ClusterSwitcher();
+ * const aggregator = new HealthAggregator(switcher);
+ *
+ * // Get aggregated health summary
+ * const summary = await aggregator.aggregateHealth();
+ * console.log(`Total clusters: ${summary.totalClusters}`);
+ * console.log(`Healthy: ${summary.healthyClusters}`);
+ * console.log(`Degraded: ${summary.degradedClusters}`);
+ * console.log(`Down: ${summary.downClusters}`);
+ *
+ * // Print formatted summary
+ * const text = await aggregator.getHealthSummaryText();
+ * console.log(text);
+ * ```
+ */
+
+/**
+ * Example Usage: Alert on Unhealthy Clusters
+ *
+ * ```typescript
+ * const switcher = new ClusterSwitcher();
+ * const aggregator = new HealthAggregator(switcher);
+ *
+ * // Check for unhealthy clusters
+ * if (await aggregator.hasUnhealthyClusters()) {
+ *   const unhealthy = await aggregator.getUnhealthyClusters();
+ *   for (const cluster of unhealthy) {
+ *     console.error(`⚠ ${cluster.clusterName} is ${cluster.status}`);
+ *     console.error(`  Under-replicated: ${cluster.underReplicatedPartitions}`);
+ *     console.error(`  Offline: ${cluster.offlinePartitions}`);
+ *   }
+ *   // Send alert to PagerDuty, Slack, etc.
+ * }
+ * ```
+ */
+
+/**
+ * Example Usage: Scheduled Health Monitoring
+ *
+ * ```typescript
+ * const switcher = new ClusterSwitcher();
+ * const aggregator = new HealthAggregator(switcher);
+ *
+ * // Run health check every 5 minutes
+ * setInterval(async () => {
+ *   const summary = await aggregator.aggregateHealth();
+ *
+ *   // Export to Prometheus (push gateway)
+ *   pushToPrometheus('kafka_clusters_total', summary.totalClusters);
+ *   pushToPrometheus('kafka_clusters_healthy', summary.healthyClusters);
+ *   pushToPrometheus('kafka_clusters_degraded', summary.degradedClusters);
+ *   pushToPrometheus('kafka_clusters_down', summary.downClusters);
+ *
+ *   // Alert if any cluster is down
+ *   if (summary.downClusters > 0) {
+ *     sendSlackAlert(`⚠ ${summary.downClusters} Kafka cluster(s) down!`);
+ *   }
+ * }, 5 * 60 * 1000); // 5 minutes
+ * ```
+ */
+
+export default HealthAggregator;

package/plugins/specweave-kafka/lib/observability/opentelemetry-kafka.js

@@ -0,0 +1,209 @@
+import {
+  trace,
+  context,
+  SpanKind,
+  SpanStatusCode,
+  propagation
+} from "@opentelemetry/api";
+const KafkaAttributes = {
+  MESSAGING_SYSTEM: "messaging.system",
+  MESSAGING_DESTINATION: "messaging.destination",
+  MESSAGING_DESTINATION_KIND: "messaging.destination_kind",
+  MESSAGING_OPERATION: "messaging.operation",
+  MESSAGING_KAFKA_MESSAGE_KEY: "messaging.kafka.message_key",
+  MESSAGING_KAFKA_PARTITION: "messaging.kafka.partition",
+  MESSAGING_KAFKA_OFFSET: "messaging.kafka.offset",
+  MESSAGING_KAFKA_TOMBSTONE: "messaging.kafka.tombstone",
+  MESSAGING_KAFKA_CLIENT_ID: "messaging.kafka.client_id",
+  MESSAGING_KAFKA_CONSUMER_GROUP: "messaging.kafka.consumer.group"
+};
+const TRACEPARENT_HEADER = "traceparent";
+const TRACESTATE_HEADER = "tracestate";
+class KafkaProducerTracing {
+  constructor(config = {}) {
+    this.config = {
+      tracerName: config.tracerName || "kafka-producer",
+      capturePayloads: config.capturePayloads || false,
+      captureKeys: config.captureKeys || true,
+      maxPayloadSize: config.maxPayloadSize || 1024
+    };
+    this.tracer = trace.getTracer(this.config.tracerName);
+  }
+  /**
+   * Instrument a producer.send() call with distributed tracing
+   */
+  async traceSend(producer, topic, messages, clientId) {
+    return this.tracer.startActiveSpan(
+      `${topic} send`,
+      {
+        kind: SpanKind.PRODUCER,
+        attributes: {
+          [KafkaAttributes.MESSAGING_SYSTEM]: "kafka",
+          [KafkaAttributes.MESSAGING_DESTINATION]: topic,
+          [KafkaAttributes.MESSAGING_DESTINATION_KIND]: "topic",
+          [KafkaAttributes.MESSAGING_OPERATION]: "send",
+          [KafkaAttributes.MESSAGING_KAFKA_CLIENT_ID]: clientId || "unknown",
+          "messaging.batch.message_count": messages.length
+        }
+      },
+      async (span) => {
+        try {
+          const instrumentedMessages = messages.map((msg) => {
+            const headers = msg.headers || {};
+            propagation.inject(context.active(), headers, {
+              set: (carrier, key, value) => {
+                carrier[key] = value;
+              }
+            });
+            if (this.config.captureKeys && msg.key) {
+              span.setAttribute(
+                KafkaAttributes.MESSAGING_KAFKA_MESSAGE_KEY,
+                msg.key.toString()
+              );
+            }
+            if (this.config.capturePayloads && msg.value) {
+              const payload = msg.value.toString();
+              const truncated = payload.length > this.config.maxPayloadSize ? payload.substring(0, this.config.maxPayloadSize) + "..." : payload;
+              span.setAttribute("messaging.message.payload", truncated);
+            }
+            return { ...msg, headers };
+          });
+          const result = await producer.send({
+            topic,
+            messages: instrumentedMessages
+          });
+          span.setStatus({ code: SpanStatusCode.OK });
+          result.forEach((metadata, index) => {
+            span.addEvent("message.sent", {
+              [KafkaAttributes.MESSAGING_KAFKA_PARTITION]: metadata.partition,
+              [KafkaAttributes.MESSAGING_KAFKA_OFFSET]: metadata.offset?.toString() || "unknown",
+              "message.index": index
+            });
+          });
+          return result;
+        } catch (error) {
+          span.setStatus({
+            code: SpanStatusCode.ERROR,
+            message: error instanceof Error ? error.message : String(error)
+          });
+          span.recordException(error);
+          throw error;
+        } finally {
+          span.end();
+        }
+      }
+    );
+  }
+}
+class KafkaConsumerTracing {
+  constructor(config = {}) {
+    this.config = {
+      tracerName: config.tracerName || "kafka-consumer",
+      capturePayloads: config.capturePayloads || false,
+      captureKeys: config.captureKeys || true,
+      maxPayloadSize: config.maxPayloadSize || 1024
+    };
+    this.tracer = trace.getTracer(this.config.tracerName);
+  }
+  /**
+   * Instrument a message consumption with distributed tracing
+   */
+  async traceMessage(topic, partition, message, consumerGroup, handler) {
+    const parentContext = propagation.extract(
+      context.active(),
+      message.headers || {},
+      {
+        get: (carrier, key) => {
+          const value = carrier[key];
+          return Array.isArray(value) ? value[0]?.toString() : value?.toString();
+        },
+        keys: (carrier) => Object.keys(carrier)
+      }
+    );
+    return this.tracer.startActiveSpan(
+      `${topic} receive`,
+      {
+        kind: SpanKind.CONSUMER,
+        attributes: {
+          [KafkaAttributes.MESSAGING_SYSTEM]: "kafka",
+          [KafkaAttributes.MESSAGING_DESTINATION]: topic,
+          [KafkaAttributes.MESSAGING_DESTINATION_KIND]: "topic",
+          [KafkaAttributes.MESSAGING_OPERATION]: "receive",
+          [KafkaAttributes.MESSAGING_KAFKA_PARTITION]: partition,
+          [KafkaAttributes.MESSAGING_KAFKA_OFFSET]: message.offset,
+          [KafkaAttributes.MESSAGING_KAFKA_CONSUMER_GROUP]: consumerGroup,
+          [KafkaAttributes.MESSAGING_KAFKA_TOMBSTONE]: message.value === null
+        }
+      },
+      parentContext,
+      async (span) => {
+        try {
+          if (this.config.captureKeys && message.key) {
+            span.setAttribute(
+              KafkaAttributes.MESSAGING_KAFKA_MESSAGE_KEY,
+              message.key.toString()
+            );
+          }
+          if (this.config.capturePayloads && message.value) {
+            const payload = message.value.toString();
+            const truncated = payload.length > this.config.maxPayloadSize ? payload.substring(0, this.config.maxPayloadSize) + "..." : payload;
+            span.setAttribute("messaging.message.payload", truncated);
+          }
+          const result = await handler(context.active());
+          span.setStatus({ code: SpanStatusCode.OK });
+          return result;
+        } catch (error) {
+          span.setStatus({
+            code: SpanStatusCode.ERROR,
+            message: error instanceof Error ? error.message : String(error)
+          });
+          span.recordException(error);
+          throw error;
+        } finally {
+          span.end();
+        }
+      }
+    );
+  }
+  /**
+   * Create a processing span for custom operations within message handler
+   */
+  createProcessingSpan(name, operation, callback) {
+    return this.tracer.startActiveSpan(
+      name,
+      {
+        kind: SpanKind.INTERNAL,
+        attributes: {
+          "operation.name": operation
+        }
+      },
+      async (span) => {
+        try {
+          const result = await callback(span);
+          span.setStatus({ code: SpanStatusCode.OK });
+          return result;
+        } catch (error) {
+          span.setStatus({
+            code: SpanStatusCode.ERROR,
+            message: error instanceof Error ? error.message : String(error)
+          });
+          span.recordException(error);
+          throw error;
+        } finally {
+          span.end();
+        }
+      }
+    );
+  }
+}
+var opentelemetry_kafka_default = {
+  KafkaProducerTracing,
+  KafkaConsumerTracing,
+  KafkaAttributes
+};
+export {
+  KafkaAttributes,
+  KafkaConsumerTracing,
+  KafkaProducerTracing,
+  opentelemetry_kafka_default as default
+};