OPTIMASI BUILD DAN DEPLOYMENT APLIKASI BUN-FASTIFY DAN NEXT.JS MENGGUNAKAN DOCKER MULTI-STAGE BUILD PADA SISTEM AKADEMIK UMPO

Sistem Informasi Akademik (SIMAK) of Universitas Muhammadiyah Ponorogo, consisting of admin frontend, student frontend, and backend services, faces the problem of large Docker image sizes, which result in slower deployment processes and inefficient use of storage resources and bandwidth during image distribution. This study aims to optimize the Docker image sizes of these three services using a multi-stage build approach to improve the efficiency of building and deploying applications based on Bun-Fastify and Next.js. The study employed a quantitative approach by building two image versions for each service: an unoptimized version using a single-stage build and an optimized version using a multi-stage build that separates the builder and runner stages. The image size, number of layers, and build context size were then compared using Docker Engine and Portainer as management tools. The results showed that the multi-stage build approach successfully reduced the backend image size from 709 MB to 347 MB (51%), the admin frontend from 1.92 GB to 373 MB (83.5%), and the student frontend from 1.46 GB to 386 MB (71.3%). The significant reduction in frontend image sizes was achieved through the utilization of Next.js standalone output, which performs dependency tree-shaking, while the reduction in the backend was achieved by separating compiled artifacts from the source code and development dependencies. This study concludes that implementing multi-stage builds is an effective strategy for optimizing Docker image sizes in academic information systems based on Bun and Next.js.