Laravel Redis Queue Workers Crashing on Nginx: 5 Urgent Fixes to Stop Brutal Job Failures and Restore 100% Reliability in 30 Minutes

You’ve just pushed a hot‑fix to production, but seconds later your Laravel queue workers start dying like flies. The logs scream “Redis connection lost”, your API slows to a crawl, and your customers start seeing error pages. If you’re a PHP dev juggling VPS hosting, Nginx, and Redis, you know the frustration is real. This article cuts through the noise and gives you five battle‑tested fixes you can apply in under 30 minutes—no reboot, no downtime, just pure reliability.

Why This Matters

Queue workers are the backbone of any modern Laravel application—email notifications, webhook dispatches, image processing, you name it. When they crash:

• Revenue‑generating jobs are lost.
• Customer trust erodes faster than a bad CDN cache.
• Ops teams waste precious hours chasing phantom Redis timeouts.

Fixing the root cause not only restores 100% job success but also improves overall PHP‑FPM and MySQL throughput, giving you a smoother experience for both Laravel and any WordPress sites sharing the same VPS.

Common Causes of Crashy Workers

1. Supervisor misconfiguration: Workers are killed when they exceed default memory limits.
2. Redis connection limits: Too many concurrent connections exhaust the default 10,000 limit.
3. Nginx fastcgi buffers: Improper buffering causes upstream timeouts.
4. PHP‑FPM pm.max_children: Under‑provisioned children lead to request queueing.
5. Missing .env production flag: Workers run in “local” mode, disabling critical caching.

INFO: The fixes below assume you’re on Ubuntu 22.04 LTS with php8.2-fpm, nginx, and redis-server installed. Adjust package names for Debian or CentOS accordingly.

Step‑by‑Step Fix Tutorial

1️⃣ Tune Supervisor for Laravel Workers

Open your Supervisor config (usually /etc/supervisor/conf.d/laravel-worker.conf) and apply these settings:

[program:laravel-queue]
process_name=%(program_name)s_%(process_num)02d
command=php /var/www/html/artisan queue:work redis --sleep=3 --tries=3 --timeout=90
autostart=true
autorestart=true
user=www-data
numprocs=8
directory=/var/www/html
stopwaitsecs=360
stdout_logfile=/var/log/laravel/worker.log
stderr_logfile=/var/log/laravel/worker_error.log
environment=APP_ENV=production,APP_DEBUG=false
# Prevent OOM kills
memory_limit=256M

After saving, reload Supervisor:

sudo supervisorctl reread
sudo supervisorctl update
sudo supervisorctl restart laravel-queue:*

TIP: Set numprocs based on cpu cores * 2 for optimal concurrency on a typical 2 vCPU VPS.

2️⃣ Raise Redis Max Connections

Open /etc/redis/redis.conf and increase maxclients:

maxclients 20000
tcp-backlog 511
timeout 0

Restart Redis to apply:

sudo systemctl restart redis

WARNING: Do not set maxclients higher than your RAM can support. Each client consumes ~4 KB.

3️⃣ Optimize Nginx FastCGI Buffers

Add these directives inside the server block that serves your Laravel app:

location ~ \.php$ {
    include fastcgi_params;
    fastcgi_pass unix:/run/php/php8.2-fpm.sock;
    fastcgi_read_timeout 300;
    fastcgi_buffer_size 32k;
    fastcgi_buffers 8 64k;
    fastcgi_busy_buffers_size 128k;
    fastcgi_temp_file_write_size 256k;
}

Reload Nginx:

sudo nginx -t && sudo systemctl reload nginx

4️⃣ Boost PHP‑FPM pm.max_children

Edit /etc/php/8.2/fpm/pool.d/www.conf and set:

pm = dynamic
pm.max_children = 30
pm.start_servers = 6
pm.min_spare_servers = 4
pm.max_spare_servers = 12

Then restart PHP‑FPM:

sudo systemctl restart php8.2-fpm

5️⃣ Enforce Production Environment in .env

Make sure the .env file reflects the real environment:

APP_ENV=production
APP_DEBUG=false
CACHE_DRIVER=redis
QUEUE_CONNECTION=redis
REDIS_HOST=127.0.0.1
REDIS_PASSWORD=null
REDIS_PORT=6379

Clear any stale config cache:

php artisan config:clear
php artisan cache:clear
php artisan queue:restart

SUCCESS: After these changes, monitor supervisorctl status and Redis logs for 10‑15 minutes. You should see worker processes staying alive with zero “connection lost” errors.

VPS or Shared Hosting Optimization Tips

• Swap management: On low‑memory VPS, add a 1 GB swap file to prevent OOM kills.
• Uptime monitoring: Use UptimeRobot to alert on queue length spikes.
• Separate Redis instance: If you share a VPS with WordPress, consider a dedicated Redis port (e.g., 6380) for Laravel.
• Cloudflare caching: Bypass Cloudflare for /api/* calls that trigger queue jobs to reduce latency.
• Composer autoloader optimization: Run composer install --optimize-autoloader --no-dev on production.

Real World Production Example

Acme Media runs a Laravel‑powered newsletter platform on a 2 vCPU, 4 GB RAM VPS. After the Redis crash, they applied the five fixes above. Within 30 minutes they observed:

• Queue failure rate dropped from 23% to 0%.
• Average job processing time improved from 4.8 s to 1.2 s.
• CPU usage stabilized at ~35% during peak email bursts.

Before vs After Results

Metric	Before Fix	After Fix
Failed Jobs	23%	0%
Avg. Job Time	4.8 s	1.2 s
Redis CPU Load	85%	45%

Security Considerations

While tuning Redis and PHP‑FPM, never expose ports to the public internet. Add these lines to /etc/redis/redis.conf:

bind 127.0.0.1 ::1
protected-mode yes
requirepass YOUR_STRONG_PASSWORD

Restart Redis after changes. Also, keep your supervisor config files owned by root and chmod 640 to prevent tampering.

Bonus Performance Tips

• Use Horizon: Laravel Horizon gives you a UI to monitor queue health in real time.
• Batch Jobs: Group small tasks into a single job to reduce Redis round‑trips.
• Enable OPCache: Add opcache.enable=1 in php.ini for up to 15% faster script execution.
• Connection pooling: Install php-redis extension with persistent connections.
• Docker alternative: If you shift to Docker, use php-fpm and redis:alpine images with proper resource limits.

FAQ

Q: My queue still restarts after the fixes. What else can I check?

A: Look for out‑of‑memory kills in /var/log/kern.log. If they appear, increase swap or downsize memory_limit per worker.

Q: Can I use the same Redis instance for WordPress object caching?

A: Yes, but separate key prefixes (e.g., laravel: vs wp_) and allocate enough maxclients for the combined load.

Q: Do I need to adjust supervisor on shared hosting?

A: Most shared hosts don’t allow Supervisor. In that case, use Laravel’s queue:work --daemon with a cron entry running every minute.

Final Thoughts

Queue reliability is a non‑negotiable part of any SaaS or high‑traffic WordPress/Laravel hybrid. The five fixes above address the most common failure points—Supervisor limits, Redis connections, Nginx buffers, PHP‑FPM sizing, and proper environment configuration. Apply them, watch the metrics settle, and you’ll regain the confidence that your background jobs will finish on time, every time.

Need a cheap, secure VPS that already includes Nginx, PHP‑FPM, and Redis pre‑installed? Check out Hostinger’s affordable plans—they’re optimized for Laravel and WordPress, plus you’ll get a 30‑day money‑back guarantee.