Add caching ADR

doc/adr/0008-use-serverside-caching.md

@@ -0,0 +1,56 @@
+# 8. Use Server-Side Caching with NodeCache for Strapi Data Fetching
+
+**Date:** 2024-12-05
+
+## Status
+
+Accepted
+
+## Context
+
+Fetching data from Strapi frequently can result in performance
+bottlenecks and increased latency, especially when querying the same
+data repeatedly in a short time span. This can impact user experience.
+
+Previously, our approach did not involve caching, leading to:
+
+- **Repeated requests** for identical GraphQL queries, increasing load on the Strapi server.
+- **Higher response times** for end-users, particularly for frequently accessed resources.
+
+To address these issues, we evaluated multiple solutions for
+implementing caching to reduce redundant calls and improve performance.
+
+### Evaluated Options
+
+1. **Caching in Strapi with Apollo:**
+
+   - Apollo's built-in caching features were explored for use within Strapi.
+   - However, integration issues made it impractical for our needs, and it failed to provide the expected benefits.
+
+2. **External Caching Services (e.g., Redis):**
+
+   - High scalability and flexibility.
+   - Requires additional infrastructure setup and maintenance.
+   - Overhead may not be justified for our current scale.
+
+3. **In-Memory Caching with NodeCache:**
+   - Simple to integrate directly into the application code.
+   - Minimal setup with built-in TTL (time-to-live) and periodic cleanup.
+   - Well-suited for small-scale, server-side caching needs.
+
+## Decision
+
+We decided to use **NodeCache** for implementing server-side caching in our data-fetching layer.
+This approach is lightweight and integrates seamlessly with our existing architecture.
+
+Key considerations for choosing NodeCache:
+
+- **Simplicity:** Straightforward Integration.
+- **Low Overhead:** As an in-memory solution, it requires no additional external service or infrastructure.
+- **Global Caching:** We opted to cache entire responses for specific queries globally instead of caching individual fields. This is because we do not have a large variety of different requests, making global caching simpler and more effective.
+- **TTL Management:** Built-in support for automatic expiration of cached data aligns with our need for caching dynamic content with a short lifespan.
+- **Immediate Benefits:** It reduces repeated requests for identical GraphQL queries to Strapi, improving performance and decreasing server load.
+
+## Consequences
+
+We will use `node-cache` to cache requests to Strapi.