feat: add event cleanup scripts and Docker Compose setup for automated maintenance

README.md

@@ -445,6 +445,19 @@ Try the following steps:
 >   -v gitea-mirror-data:/app/data \
 >   ghcr.io/arunavo4/gitea-mirror:latest
 > ```
+>
+> For homelab/self-hosted setups, you can use the provided Docker Compose file with automatic event cleanup:
+>
+> ```bash
+> # Clone the repository
+> git clone https://github.com/arunavo4/gitea-mirror.git
+> cd gitea-mirror
+>
+> # Start the application with Docker Compose
+> docker-compose -f docker-compose.homelab.yml up -d
+> ```
+>
+> This setup includes a cron job that runs daily to clean up old events and prevent the database from growing too large.
 
 
 #### Database Maintenance
@@ -461,6 +474,19 @@ Try the following steps:
 >
 > # Reset user accounts (for development)
 > bun run reset-users
+>
+> # Clean up old events (keeps last 7 days by default)
+> bun run cleanup-events
+>
+> # Clean up old events with custom retention period (e.g., 30 days)
+> bun run cleanup-events 30
+> ```
+>
+> For automated maintenance, consider setting up a cron job to run the cleanup script periodically:
+>
+> ```bash
+> # Add this to your crontab (runs daily at 2 AM)
+> 0 2 * * * cd /path/to/gitea-mirror && bun run cleanup-events
 > ```
 
 

crontab

@@ -0,0 +1,4 @@
+# Run event cleanup daily at 2 AM
+0 2 * * * cd /app && bun run cleanup-events 30 >> /app/data/cleanup-events.log 2>&1
+
+# Empty line at the end is required for cron to work properly

docker-compose.homelab.yml

@@ -0,0 +1,38 @@
+version: '3.8'
+
+services:
+  gitea-mirror:
+    image: ghcr.io/arunavo4/gitea-mirror:latest
+    container_name: gitea-mirror
+    restart: unless-stopped
+    ports:
+      - "4321:4321"
+    volumes:
+      - gitea-mirror-data:/app/data
+      # Mount the crontab file
+      - ./crontab:/etc/cron.d/gitea-mirror-cron
+    environment:
+      - NODE_ENV=production
+      - HOST=0.0.0.0
+      - PORT=4321
+      - DATABASE_URL=sqlite://data/gitea-mirror.db
+      - DELAY=${DELAY:-3600}
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:4321/health"]
+      interval: 1m
+      timeout: 10s
+      retries: 3
+      start_period: 30s
+    # Install cron in the container and set up the cron job
+    command: >
+      sh -c "
+        apt-get update && apt-get install -y cron curl &&
+        chmod 0644 /etc/cron.d/gitea-mirror-cron &&
+        crontab /etc/cron.d/gitea-mirror-cron &&
+        service cron start &&
+        bun dist/server/entry.mjs
+      "
+
+# Define named volumes for database persistence
+volumes:
+  gitea-mirror-data:    # Database volume

package.json

@@ -18,6 +18,7 @@
     "reset-users": "bun scripts/manage-db.ts reset-users",
     "migrate-db": "bun scripts/migrate-db.ts",
     "cleanup-redis": "bun scripts/cleanup-redis.ts",
+    "cleanup-events": "bun scripts/cleanup-events.ts",
     "preview": "bunx --bun astro preview",
     "start": "bun dist/server/entry.mjs",
     "start:fresh": "bun run cleanup-db && bun run manage-db init && bun dist/server/entry.mjs",

scripts/cleanup-events.ts

@@ -0,0 +1,43 @@
+#!/usr/bin/env bun
+/**
+ * Script to clean up old events from the database
+ * This script should be run periodically (e.g., daily) to prevent the events table from growing too large
+ *
+ * Usage:
+ *   bun scripts/cleanup-events.ts [days]
+ *
+ * Where [days] is the number of days to keep events (default: 7)
+ */
+
+import { cleanupOldEvents } from "../src/lib/events";
+
+// Parse command line arguments
+const args = process.argv.slice(2);
+const daysToKeep = args.length > 0 ? parseInt(args[0], 10) : 7;
+
+if (isNaN(daysToKeep) || daysToKeep < 1) {
+  console.error("Error: Days to keep must be a positive number");
+  process.exit(1);
+}
+
+async function runCleanup() {
+  try {
+    console.log(`Starting event cleanup (retention: ${daysToKeep} days)...`);
+
+    // Call the cleanupOldEvents function from the events module
+    const result = await cleanupOldEvents(daysToKeep);
+
+    console.log(`Cleanup summary:`);
+    console.log(`- Read events deleted: ${result.readEventsDeleted}`);
+    console.log(`- Unread events deleted: ${result.unreadEventsDeleted}`);
+    console.log(`- Total events deleted: ${result.readEventsDeleted + result.unreadEventsDeleted}`);
+
+    console.log("Event cleanup completed successfully");
+  } catch (error) {
+    console.error("Error running event cleanup:", error);
+    process.exit(1);
+  }
+}
+
+// Run the cleanup
+runCleanup();

scripts/make-events-old.ts

@@ -0,0 +1,29 @@
+#!/usr/bin/env bun
+/**
+ * Script to make events appear older for testing cleanup
+ */
+
+import { db, events } from "../src/lib/db";
+
+async function makeEventsOld() {
+  try {
+    console.log("Making events appear older...");
+    
+    // Calculate a timestamp from 2 days ago
+    const oldDate = new Date();
+    oldDate.setDate(oldDate.getDate() - 2);
+    
+    // Update all events to have an older timestamp
+    const result = await db
+      .update(events)
+      .set({ createdAt: oldDate });
+    
+    console.log(`Updated ${result.changes || 0} events to appear older`);
+  } catch (error) {
+    console.error("Error updating event timestamps:", error);
+    process.exit(1);
+  }
+}
+
+// Run the function
+makeEventsOld();

scripts/mark-events-read.ts

@@ -0,0 +1,27 @@
+#!/usr/bin/env bun
+/**
+ * Script to mark all events as read
+ */
+
+import { db, events } from "../src/lib/db";
+import { eq } from "drizzle-orm";
+
+async function markEventsAsRead() {
+  try {
+    console.log("Marking all events as read...");
+    
+    // Update all events to mark them as read
+    const result = await db
+      .update(events)
+      .set({ read: true })
+      .where(eq(events.read, false));
+    
+    console.log(`Marked ${result.changes || 0} events as read`);
+  } catch (error) {
+    console.error("Error marking events as read:", error);
+    process.exit(1);
+  }
+}
+
+// Run the function
+markEventsAsRead();

src/lib/events.ts

@@ -1,6 +1,6 @@
 import { v4 as uuidv4 } from "uuid";
 import { db, events } from "./db";
-import { eq, and, gt } from "drizzle-orm";
+import { eq, and, gt, lt } from "drizzle-orm";
 
 /**
  * Publishes an event to a specific channel for a user
@@ -106,25 +106,56 @@ export async function getNewEvents({
 /**
  * Cleans up old events to prevent the database from growing too large
  * Should be called periodically (e.g., daily via a cron job)
+ *
+ * @param maxAgeInDays Number of days to keep events (default: 7)
+ * @param cleanupUnreadAfterDays Number of days after which to clean up unread events (default: 2x maxAgeInDays)
+ * @returns Object containing the number of read and unread events deleted
  */
-export async function cleanupOldEvents(maxAgeInDays: number = 7): Promise<number> {
+export async function cleanupOldEvents(
+  maxAgeInDays: number = 7,
+  cleanupUnreadAfterDays?: number
+): Promise<{ readEventsDeleted: number; unreadEventsDeleted: number }> {
   try {
+    console.log(`Cleaning up events older than ${maxAgeInDays} days...`);
+
+    // Calculate the cutoff date for read events
     const cutoffDate = new Date();
     cutoffDate.setDate(cutoffDate.getDate() - maxAgeInDays);
 
-    // Delete events older than the cutoff date
-    const result = await db
+    // Delete read events older than the cutoff date
+    const readResult = await db
       .delete(events)
       .where(
         and(
           eq(events.read, true),
-          gt(cutoffDate, events.createdAt)
+          lt(events.createdAt, cutoffDate)
         )
       );
 
-    return result.changes || 0;
+    const readEventsDeleted = readResult.changes || 0;
+    console.log(`Deleted ${readEventsDeleted} read events`);
+
+    // Calculate the cutoff date for unread events (default to 2x the retention period)
+    const unreadCutoffDate = new Date();
+    const unreadMaxAge = cleanupUnreadAfterDays || (maxAgeInDays * 2);
+    unreadCutoffDate.setDate(unreadCutoffDate.getDate() - unreadMaxAge);
+
+    // Delete unread events that are significantly older
+    const unreadResult = await db
+      .delete(events)
+      .where(
+        and(
+          eq(events.read, false),
+          lt(events.createdAt, unreadCutoffDate)
+        )
+      );
+
+    const unreadEventsDeleted = unreadResult.changes || 0;
+    console.log(`Deleted ${unreadEventsDeleted} unread events`);
+
+    return { readEventsDeleted, unreadEventsDeleted };
   } catch (error) {
     console.error("Error cleaning up old events:", error);
-    return 0;
+    return { readEventsDeleted: 0, unreadEventsDeleted: 0 };
   }
 }

src/pages/api/sse/index.ts

@@ -11,7 +11,7 @@ export const GET: APIRoute = async ({ request }) => {
 
   const channel = `mirror-status:${userId}`;
   let isClosed = false;
-  const POLL_INTERVAL = 2000; // Poll every 2 seconds
+  const POLL_INTERVAL = 5000; // Poll every 5 seconds (reduced from 2 seconds for low-traffic usage)
 
   const stream = new ReadableStream({
     start(controller) {