feat: add retry mechanism with exponential backoff for download failures

src/main/services/playwright-browser/download-service.ts

@@ -88,6 +88,92 @@ export class DownloadService {
     this.s3Client = this.config.s3Client
   }
 
+  /**
+   * Execute an operation with retry logic using exponential backoff
+   * @param operation - The async operation to execute
+   * @param context - Description of the operation for logging
+   * @param maxRetries - Maximum number of retry attempts (default: 3)
+   * @returns The result of the operation
+   */
+  private async withRetry<T>(
+    operation: () => Promise<T>,
+    context: string,
+    maxRetries = 3
+  ): Promise<T> {
+    let lastError: Error | undefined
+
+    for (let attempt = 1; attempt <= maxRetries; attempt++) {
+      try {
+        return await operation()
+      } catch (error) {
+        lastError = error instanceof Error ? error : new Error(String(error))
+
+        // Check if error is retryable
+        const isRetryable = this.isRetryableError(lastError)
+        if (!isRetryable || attempt === maxRetries) {
+          break
+        }
+
+        // Exponential backoff: 1s, 2s, 4s
+        const delay = Math.pow(2, attempt - 1) * 1000
+        log.warn(`${context} failed (attempt ${attempt}/${maxRetries}), retrying in ${delay}ms`, {
+          error: lastError.message
+        })
+
+        await new Promise((resolve) => setTimeout(resolve, delay))
+      }
+    }
+
+    throw lastError
+  }
+
+  /**
+   * Determine if an error is retryable based on error type and message
+   * @param error - The error to classify
+   * @returns true if the error should trigger a retry
+   */
+  private isRetryableError(error: Error): boolean {
+    const message = error.message.toLowerCase()
+    const code = (error as any).code?.toLowerCase() || ''
+
+    // Retryable: network errors, timeouts
+    const retryablePatterns = [
+      'etimedout',
+      'econnreset',
+      'timeout',
+      'network',
+      'socket hang up',
+      'connection reset'
+    ]
+
+    // Not retryable: S3 errors, validation errors
+    const nonRetryablePatterns = [
+      '404',
+      '403',
+      'not found',
+      'access denied',
+      'invalid',
+      'validation'
+    ]
+
+    // Check non-retryable first
+    for (const pattern of nonRetryablePatterns) {
+      if (message.includes(pattern) || code.includes(pattern)) {
+        return false
+      }
+    }
+
+    // Check retryable
+    for (const pattern of retryablePatterns) {
+      if (message.includes(pattern) || code.includes(pattern)) {
+        return true
+      }
+    }
+
+    // Default: don't retry unknown errors
+    return false
+  }
+
   /**
    * List all S3 objects under the configured prefix
    * Filters to only chromium-* folders
@@ -130,31 +216,33 @@ export class DownloadService {
   }
 
   /**
-   * Download a single file from S3
+   * Download a single file from S3 with retry logic
    */
   async downloadFile(key: string, destPath: string): Promise<void> {
-    log.debug('Downloading file', { key, destPath })
+    await this.withRetry(async () => {
+      log.debug('Downloading file', { key, destPath })
 
-    await fs.promises.mkdir(path.dirname(destPath), { recursive: true })
+      await fs.promises.mkdir(path.dirname(destPath), { recursive: true })
 
-    const response = await this.s3Client.send(
-      new GetObjectCommand({
-        Bucket: this.config.bucket,
-        Key: key
+      const response = await this.s3Client.send(
+        new GetObjectCommand({
+          Bucket: this.config.bucket,
+          Key: key
+        })
+      )
+
+      const output = fs.createWriteStream(destPath)
+      const body = response.Body as NodeJS.ReadableStream
+
+      await new Promise<void>((resolve, reject) => {
+        body.on('error', reject)
+        output.on('error', reject)
+        output.on('finish', resolve)
+        body.pipe(output)
       })
-    )
 
-    const output = fs.createWriteStream(destPath)
-    const body = response.Body as NodeJS.ReadableStream
-
-    await new Promise<void>((resolve, reject) => {
-      body.on('error', reject)
-      output.on('error', reject)
-      output.on('finish', resolve)
-      body.pipe(output)
-    })
-
-    log.debug('File downloaded', { key, destPath })
+      log.debug('File downloaded', { key, destPath })
+    }, `Download ${key}`)
   }
 
   /**