Addressing Occasional Connection and Job Processing Timeouts in a Laravel-MariaDB-Redis Setup

Managing a web application that relies on Laravel, MariaDB, and Redis can sometimes present unforeseen challenges, especially when dealing with intermittent connection issues and job processing delays. If you’re running a real estate listings platform with automated offer imports from various CRM systems, maintaining smooth operation is critical—even during low traffic periods.

Understanding Your Setup

Your application resides on a single VPS with specifications of 8 CPU cores and 16 GB RAM, hosting the Laravel app, MariaDB, and Redis. Your current traffic is modest—averaging around 20 visits per minute, with occasional spikes driven by bots. During these bursts, a handful of connections might cling to the server momentarily.

While planning for future microservices architecture, you’re mindful of keeping costs low, which influences your current configuration.

The Issue at Hand

Every few days, certain background jobs exceed your 600-second timeout, despite typically completing within half a second. Additionally, your server intermittently experiences connection timeouts when making outbound requests to a headless WordPress instance on the same VPS—likely related to database or network connectivity issues.

Despite thorough research, configurations—including PHP pools, MariaDB connection limits, and slow query logs—haven’t yielded a resolution. Your logs lack concrete clues, with only indications of functions that became “stuck.”

Potential Causes and Troubleshooting Strategies

1. Connection Pooling and Limits
   Ensure your PHP-FPM pools are adequately configured to handle concurrent requests without exhausting resources. Since your setup doesn’t appear to hit maximum PHP pool capacities, the bottleneck may lie elsewhere.

2. MariaDB Configuration
   While you haven’t saturated max connections, verify other settings such as wait_timeout and interactive_timeout. MariaDB might be closing idle connections prematurely, affecting ongoing transactions.

3. Network and Inter-Process Communication
   Given that API requests to the WordPress instance sometimes time out, check server network health. Is there any packet loss, high latency, or resource contention? Use tools like ping, traceroute, or network monitoring utilities to assess stability.

4. Resource Utilization and Server Load
   Monitor CPU, memory, and disk I/O during normal operation and during timeout events. Over-utilization might hinder database responses or network requests.

5. Redis and Caching
   Ensure Redis is configured correctly and isn’t experiencing memory pressure or eviction issues