Implementasi Chatbot Berbasis Web Untuk Otomatisasi Deploymnet Aplikasi Pada Virtual Private Server

Ramadhani, Rizki (2026) Implementasi Chatbot Berbasis Web Untuk Otomatisasi Deploymnet Aplikasi Pada Virtual Private Server. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5027221013-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (22MB) | Request a copy

Abstract

Proses deployment aplikasi secara manual rentan terhadap kesalahan manusia, sementara implementasi pipeline CI/CD sering kali dianggap terlalu kompleks untuk diterapkan pada proyek skala kecil hingga menengah. Penelitian ini mengusulkan solusi berupa chatbot berbasis web yang diintegrasikan dengan kerangka kerja Agentic AI OpenClaw untuk mengotomatisasi deployment aplikasi berbasis kontainer Docker pada Virtual Private Server (VPS). Penelitian bertujuan mengevaluasi dan membandingkan kinerja tiga arsitektur Large Language Model (LLM) sebagai komponen kognitif agen, yaitu DeepSeek, GLM, dan model lokal Qwen. Pengujian komparatif dilakukan terhadap sepuluh repositori kode dengan variasi tumpukan teknologi dan tingkat kompleksitas teknis yang beragam. Hasil penelitian menunjukkan bahwa agen cerdas mampu mengeksekusi seluruh tahapan deployment secara otonom, mulai dari kloning repositori, analisis arsitektur, penyusunan skrip, hingga peluncuran aplikasi. Analisis komparatif menunjukkan bahwa DeepSeek merupakan model dengan kinerja terbaik, mencapai tingkat keberhasilan 100% dan efisiensi waktu pemrosesan tertinggi, dengan pengurangan rata-rata durasi komputasi sebesar 39,4% dibandingkan GLM dan 66,1% dibandingkan Qwen pada skenario yang setara. Kedua model berbasis cloud juga menunjukkan kemampuan adaptasi dan self-healing yang baik ketika menghadapi kesalahan pada terminal. Sebaliknya, model lokal Qwen memperlihatkan keterbatasan dalam kemampuan penalaran dengan tingkat keberhasilan parsial sebesar 60% serta gagal menangani anomali pada arsitektur dependensi yang kompleks. Berdasarkan hasil tersebut, arsitektur DeepSeek menjadi solusi kecerdasan buatan yang paling tangguh dan direkomendasikan untuk mendukung beban kerja otomatisasi infrastruktur dalam paradigma Agentic Engineering.
==================================================================================================================================
Manual application deployment is prone to human error, while implementing CI/CD pipelines is often considered too complex for small- and medium-scale projects. This study proposes a web-based chatbot integrated with the OpenClaw Agentic AI framework to automate the deployment of Docker container-based applications on Virtual Private Servers (VPS). The research aims to evaluate and compare the performance of three Large Language Model (LLM) architectures as the agent's cognitive component: DeepSeek, GLM, and the local Qwen model. Comparative experiments were conducted using ten code repositories with diverse technology stacks and varying levels of technical complexity. The results demonstrate that the intelligent agent successfully executed the entire deployment workflow autonomously, including repository cloning, architecture analysis, script generation, and application deployment. Comparative analysis indicates that DeepSeek achieved the best overall performance, obtaining a 100% success rate and the highest processing efficiency by reducing average computation time by 39.4% compared with GLM and 66.1% compared with Qwen under equivalent scenarios. Both cloud-based models also exhibited advanced adaptability and self-healing capabilities when handling terminal errors. In contrast, the local Qwen model showed limitations in reasoning ability, achieving only a 60% partial success rate and failing to resolve anomalies in complex dependency architectures. These findings indicate that the DeepSeek architecture is the most robust artificial intelligence solution and is the recommended choice for supporting infrastructure automation workloads within the Agentic Engineering paradigm.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Agentic AI, Chatbot, Deployment, Docker, Large Language Model, OpenClaw, DeepSeek, GLM, Qwen. Agentic AI, Chatbot, Deployment, Docker, Large Language Model, OpenClaw, DeepSeek, GLM, Qwen.
Subjects:	T Technology > T Technology (General) > T58.64 Information resources management
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Rizki Ramadhani
Date Deposited:	20 Jul 2026 06:00
Last Modified:	20 Jul 2026 06:00
URI:	http://repository.its.ac.id/id/eprint/135470

Actions (login required)

View Item