fix: improve worker error handling and reliability

1. Distinguish failure reasons: inspect error messages to label skipped
   packages as DOWNLOAD_FAILED, UPLOAD_FAILED, or EXTRACT_FAILED
   instead of catch-all DOWNLOAD_FAILED.

2. Detect orphaned uploads: before uploading, check if the same content
   hash already has a successful upload on the destination channel. Reuse
   the existing message ID instead of re-uploading (prevents duplicates
   when worker crashed between upload and DB write).

3. Increase timeouts: download from max(5min, GB*10min) to
   max(10min, GB*15min), upload from GB*10min to GB*15min.
   Prevents premature timeouts on slow connections.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

worker/src/db/queries.ts

@@ -62,6 +62,18 @@ export async function packageExistsByHash(contentHash: string) {
   return pkg !== null;
 }
 
+/**
+ * Find an already-uploaded package by content hash.
+ * Used to detect orphaned uploads — files that reached Telegram
+ * but whose package record was created from a previous successful run.
+ */
+export async function getUploadedPackageByHash(contentHash: string) {
+  return db.package.findFirst({
+    where: { contentHash, destMessageId: { not: null }, destChannelId: { not: null } },
+    select: { destChannelId: true, destMessageId: true },
+  });
+}
+
 /**
  * Check if a package already exists for a given source message ID
  * AND was successfully uploaded to the destination (destMessageId is set).

worker/src/tdlib/download.ts

@@ -354,10 +354,10 @@ export async function downloadFile(
     let lastLoggedPercent = 0;
     let settled = false;
 
-    // Timeout: 10 minutes per GB, minimum 5 minutes
+    // Timeout: 15 minutes per GB, minimum 10 minutes
     const timeoutMs = Math.max(
-      5 * 60_000,
-      (totalBytes / (1024 * 1024 * 1024)) * 10 * 60_000
+      10 * 60_000,
+      (totalBytes / (1024 * 1024 * 1024)) * 15 * 60_000
     );
     const timer = setTimeout(() => {
       if (!settled) {

worker/src/upload/channel.ts

@@ -94,10 +94,10 @@ async function sendAndWaitForUpload(
     let lastLoggedPercent = 0;
     let tempMsgId: number | null = null;
 
-    // Timeout: 10 minutes per GB, minimum 10 minutes
+    // Timeout: 15 minutes per GB, minimum 10 minutes
     const timeoutMs = Math.max(
       10 * 60_000,
-      (fileSizeMB / 1024) * 10 * 60_000
+      (fileSizeMB / 1024) * 15 * 60_000
     );
 
     const timer = setTimeout(() => {

worker/src/worker.ts

@@ -26,6 +26,7 @@ import {
   getExistingChannelsByTelegramId,
   getAccountById,
   deleteOrphanedPackageByHash,
+  getUploadedPackageByHash,
   upsertSkippedPackage,
   deleteSkippedPackage,
 } from "./db/queries.js";
@@ -643,6 +644,20 @@ export async function runWorkerForAccount(
   }
 }
 
+/**
+ * Infer the SkipReason from an error message so the UI shows the correct badge.
+ */
+function inferSkipReason(errMsg: string): "DOWNLOAD_FAILED" | "UPLOAD_FAILED" | "EXTRACT_FAILED" {
+  const lower = errMsg.toLowerCase();
+  if (lower.includes("upload") || lower.includes("too many requests") || lower.includes("retry after") || lower.includes("send")) {
+    return "UPLOAD_FAILED";
+  }
+  if (lower.includes("extract") || lower.includes("metadata") || lower.includes("central directory") || lower.includes("archive")) {
+    return "EXTRACT_FAILED";
+  }
+  return "DOWNLOAD_FAILED";
+}
+
 /**
  * Process a scan result through the archive pipeline:
  * group → download → hash → dedup → metadata → split → upload → preview → index.
@@ -737,11 +752,12 @@ async function processArchiveSets(
       try {
         const archiveSet = archiveSets[setIdx];
         const totalSize = archiveSet.parts.reduce((sum, p) => sum + p.fileSize, 0n);
+        const errMsg = setErr instanceof Error ? setErr.message : String(setErr);
         await upsertSkippedPackage({
           fileName: archiveSet.parts[0].fileName,
           fileSize: totalSize,
-          reason: "DOWNLOAD_FAILED",
-          errorMessage: setErr instanceof Error ? setErr.message : String(setErr),
+          reason: inferSkipReason(errMsg),
+          errorMessage: errMsg,
           sourceChannelId: ctx.channel.id,
           sourceMessageId: archiveSet.parts[0].id,
           sourceTopicId: ctx.sourceTopicId,
@@ -1017,6 +1033,18 @@ async function processOneArchiveSet(
     }
 
     // ── Uploading ──
+    // Check if a prior run already uploaded this file (orphaned upload scenario:
+    // file reached Telegram but DB write failed or worker crashed before indexing)
+    const existingUpload = await getUploadedPackageByHash(contentHash);
+    let destResult: { messageId: bigint };
+
+    if (existingUpload && existingUpload.destMessageId) {
+      accountLog.info(
+        { fileName: archiveName, destMessageId: Number(existingUpload.destMessageId) },
+        "Reusing existing upload (file already on destination channel)"
+      );
+      destResult = { messageId: existingUpload.destMessageId };
+    } else {
       const uploadLabel = uploadPaths.length > 1
         ? ` (${uploadPaths.length} parts)`
         : "";
@@ -1029,11 +1057,12 @@ async function processOneArchiveSet(
         totalFiles: totalSets,
       });
 
-    const destResult = await uploadToChannel(
+      destResult = await uploadToChannel(
         client,
         destChannelTelegramId,
         uploadPaths
       );
+    }
 
     // ── Preview thumbnail ──
     let previewData: Buffer | null = null;