I'm new to deploying applications and feel stuck in a frustrating cycle. Whenever I find an error, I edit my Dockerfile or Compose configuration, push the changes to GitHub, wait around 10 minutes for the image to build, then manually pull and run the new image on my server. Often I end up waiting all that time only to discover another error. Is this a normal workflow, or am I approaching development and deployment the wrong way?
2 Answers
You could also look into a lightweight deployment tool such as Haloy. It may help automate pushing and restarting the application instead of requiring you to pull and run every new image manually, though improving the local development loop should still be the first priority.
That feedback loop is much slower than it needs to be. First, measure how long each step takes so you can identify the real bottleneck. For day-to-day development, use a local setup where your source code is mounted into a runtime container, so you can test changes without rebuilding and pushing an image every time. Save image builds and server deployment for changes that actually need to be tested in production-like conditions. When you do need to build, make sure Docker layer caching is working and keep builds local when possible. Your Dockerfile, Compose file, and CI workflow would reveal more specific improvements.
That makes sense. I’m starting to see that building the image remotely for every small change was probably unnecessary and was adding most of the delay. I’ll look into a local development setup and Docker caching.