Noah Kawaguchi

Project maintained by noahkawaguchi Hosted on GitHub Pages — Theme by mattgraham

Project Post Week 8

Last week, I finished all of the tasks I planned to get in before the deployment phase.

I finished setting up automated database seeding, which had proven to be more involved than expected, but is now solidly automated.
I added row locking for transactions that read a row, make a business decision, and then write a different row. Previously, it would have been possible for a concurrent writer to change the row that was read between the transaction’s read and write steps.
I set up basic structured logging. I decided to go with the relatively simpler crates (libraries) log and env_logger based on the project’s current scale and needs. In the future, I could switch to the more involved tracing crate if necessary.
I added graceful shutdown to the backend server. Previously, it was just stopped immediately with Ctrl+C and exited with a nonzero status. Now, it cleans up and exits normally after allowing pending requests to finish. Specifically, this is triggered by receiving SIGINT or SIGTERM on Unix or Ctrl+C on non-Unix (although it will likely only be run on Unix).
I updated and upgraded dependencies. In the backend, this involved a small modification to the code after a breaking change. In the frontend, one of the ESLint helper packages could not be upgraded because of incompatibilities in the config file syntax, but this is only used in linting, so is of very low concern. Other than that, all dependencies are now at their latest versions, and all tests pass.
I also cleaned up various smaller details that did not affect core functionality but were minor changes I wanted to get in before deployment.

This week, I will be starting into the deployment phase. The frontend deployment (GitHub Pages) will be relatively simple and is something I have done multiple times before, so I will be specifically focusing on backend deployment. Overall, I plan to Dockerize the backend server (which uses PostgreSQL) and deploy it on AWS. After some preliminary research, it appears that this will be quite involved and require connecting many different AWS offerings, even when trying to keep things simple. However, this will be a good learning experience. If the only goal was to get my app deployed, it would not make sense to try to use AWS, but I would like to use this as an opportunity to get a better sense of the kinds of DevOps practices and tools that are used in industry. I have some experience with both Docker and AWS, but not in the way that I plan to use them in this project. Therefore, the specific focus for this week will be Docker.

There are not any completely blocking impediments this week, but as I mentioned, I am moving into territory with more unfamiliar tools and will need to spend a bit more time directly learning compared to some other weeks. I also have another class that has a lab due, a project due, and an exam next week. However, I still expect to be able to make good progress on this project.

I think my process last week worked well. I have definitely been making improvements and have been feeling more confident with the workflow of using GitHub Projects (kanban board), issues, pull requests, etc. Moving into the deployment phase will involve a different kind of work than working on the code itself, so I will strive to keep the momentum going and adapt my process accordingly.