Troubleshooting Intermittent Connection and Job Timeout Issues in a Laravel, MariaDB, and Redis Setup

Optimizing a Laravel Real Estate Listings Application Facing Sporadic Connection and Processing Delays

Introduction

Managing a Laravel-based real estate listings platform involves balancing performance, reliability, and cost-efficiency. In my current setup, the application automates offer imports through multiple external CRMs, with the entire stack—including the application, database, and cache—hosted on a single VPS. Despite relatively light traffic, I encounter occasional connection timeout issues and job processing delays that disrupt smooth operations. I’m reaching out for experienced insights and potential solutions.

System Overview

Server Environment:
– VPS with 8 CPU cores and 16 GB RAM
– All components (Laravel app, MariaDB, Redis, and headless WordPress site) reside on the same server
– Traffic is modest—up to 20 visits per minute, with occasional spikes due to bots (~10-15 connections per second)

Current Architecture:
– The app is designed to scale into micro-services in the future, but for now, cost constraints dictate a monolithic setup
– Automated offer imports trigger background jobs processed within a 600-second timeout window; however, these jobs usually complete in approximately 0.5 seconds
– The application makes periodic API requests to a local WordPress instance to retrieve blog content, which sometimes results in curl connection timeouts

The Problem

Despite the low load, I experience infrequent connection timeouts to the database and external API calls. Notably:
– Background jobs occasionally hit the 600-second processing limit
– CURL requests to the local WordPress instance intermittently fail with connection timeouts

These symptoms suggest potential instability in database connectivity or resource limitations affecting sustained connections. I’ve attempted various standard troubleshooting steps with no definitive resolution.

What I’ve Tried

• PHP Pool Configuration: Ensured PHP-FPM pools are adequately configured; no exhaustion issues observed
• MariaDB Settings: Max connections not saturated; timeout settings extended to 12 hours to prevent premature disconnections
• Logging: Enabled slow query logs—no slow queries or anomalies identified; logs show normal execution times with no obvious bottlenecks

Areas for Possible Focus
– Network stability between processes on the same VPS, especially under burst traffic
– Resource contention or limitations affecting persistent connections
– External API call handling and potential network latency
– Redis connection stability (if used for caching/session management)

Request for Guidance