Inavord Architecture - A Template for Success

Vigil Journey

The first quarter of 2020 is over and I have yet to write a single blog post. I have been tinkering away constantly, but with the change of employment I haven’t really made much progress. Lots of brain power has been spent on what I want to do, but I haven’t put together the keyboard time to build anything of substance nor in writing about it. Time to get things pushed along, though.

Inavord Architecture

I have always wanted to have some kind of a solution template from which I can build future projects. The goal being to take out the repetative overhead that projects require and to based it all on one, highly opinionated stack.

  • Vuejs 3.0
    • Typescript
    • SCSS
    • Axios (seems to be the community default)
    • Vuex (does Vue3 need a state manager?)
    • Pick an i18n framework
    • Pick a unit and end-to-end testing framework
  • Azure Services (development and deployment)
    • Visual Studio Code (maybe Visual Studio Online)
    • Github (git repo, issues, actions, project board, wiki)
    • API Management (developer portal)
  • Azure Services (solution integration)
    • Active Directory B2C (see this article)
    • Key Vault
    • Functions (loosely, one app for anonymous endpoints, one app for authenticated endpoints)
    • API Management (provisioning, securing)
    • Event Grid
    • Cosmos Db
    • Storage (static website hosting)

In an ideal situation, I would have scripts that can provision resources, push configuration settings, upload code, run tests, and launch the site. The perfect situation would allow a user to fork the repo, follow a short step-by-step tutorial, and have a fully functional solution.

What (Almost) All Projects Need

All of my projects seem to start with a basic header/body/footer template, a pair of early pages (Home and “stuff”), and some test data to render. I’ll quickly get to a point of needing to persist data to a centralized storage medium, have user authentication and authorization, testing frameworks and routines, and (worst of all) remember all the tutorials and hacks I needed to follow to get anything working. My hope is that by putting together a base framework/architecture, I can minimize the ramp-up time to creating these new projects. I have lots of ideas for little apps I want to build; I just can’t seem to get over that initial overhead involved with getting the common features created.

What Inavord Won’t Include

I am not trying to make this a pluggable architecture where I expose all kinds of hooks, extensibility, and package support. I just want a base set of code that I can reuse. I know this won’t scale very well if I have dozens of projects and I find something I need to fix. If I get to that point of scale, I’ll happily refactor everything. I am not going to get hung up on making it that adaptable. That’s future David’s problem.

Getting Started?

To be frank, I have no clue where to start. By brain is so scattered with ideas and hopes and dreams that I just want the results to magically appear - or to have teams of people to do the work where I can directly them to build individual parts. There is so much for me to learn about that it becomes overwhelming sometimes (all the time). Even just thinking about related steps like “deploy environment” and “manage secrets” - which comes first? Do I learn how to deploy an environment (including the Azure Key Vault instance) and then figure out how to securely store the environment secrets? Or do I store secrets in Key Vault (including app settings for production environment) and then figure out how to read from the store to deploy the environment? And why can’t I just work on something with early tangiable results like the front-end?

Years ago, I started the project and got absolutely nowhere with it. I am going to tear it all down and restart it from scratch with the latest framework decisions. Follow along at drovani/inavord.