The dynamic and ever-evolving landscape of the Internet has introduced numerous vulnerabilities that threaten the security of online systems. While cyber attacks grow increasingly sophisticated, current detection systems typically handle each type of attack independently and lack a centralized approach for comprehensive threat monitoring. We propose a centralized multi-model deep learning framework capable of simultaneously monitoring and detecting multiple web threats, specifically DDoS attacks and Code-Injection vulnerabilities. Our model, trained on the ISCXIDS2012 dataset, achieves an impressive accuracy of 97.48%, demonstrating its effectiveness in concurrent threat detection. This unified approach significantly reduces the complexity and resource overhead associated with maintaining multiple independent detection systems. Looking ahead, the framework could be extended to incorporate detection capabilities for emerging threat vectors such as zero-day exploits. Implementing real-time adaptive learning mechanisms could further enable the system to evolve with new attack patterns without requiring complete retraining.