Merge pull request #21 from dxw/adr-0003

Add ADR for storing repo data as JSON blob

doc/architecture/decisions/0002-keep-dependencies-to-a-minimum.md

@@ -15,7 +15,7 @@ To begin with the application is a simple dashboard.
 
 ## Decision
 
-We will keep the dependencies of this project to an absolute medium.
+We will keep the dependencies of this project to an absolute minimum.
 
 ## Consequences
 

doc/architecture/decisions/0003-store-data-in-json-blob.md

@@ -0,0 +1,33 @@
+# 3. Store repository data in JSON blob
+
+Date: 2024-09-10
+
+## Status
+
+Accepted
+
+## Context
+
+- The data we need to make a usable dashboard cannot be constructed from a single API call.
+- At a minimum to display the number of open PRs for a repo we will need to GET an organisations repos and then GET on each repo.
+- If we then want to sort this data (EG to show which repos have the most open PRs) we will then need to call all of these endpoints again for each repo.
+- Therefore we will need to store this data somewhere
+- The project is in a nascent stage
+- Per [ADR-0002](./0002-keep-dependencies-to-a-minimum.md) we are keeping dependencies to a minimum
+- A decision on a database and a database structure at this point could constrain future development
+- The data used in the application is read-only
+
+
+## Decision
+
+We will store the data the application needs in JSON.
+
+## Consequences
+
+- The data can be persisted between deployments. This wouldn't be possible if stored in-memory.
+- We don't have to describe a rigid database structure.
+- We don't have to add any dependencies.
+- Initial deployments will be more simple
+- We won't be able to query the data using SQL
+- It may be harder to keep the data up to date as we won't have a database engine to lean on. We will have to perform updates manually.
+- We should revist this decision once the project is off the ground.