When the cache TTL expires, the cache will not be refreshed until the next user visits the site, because refreshing the cache is triggered by a request. This means that this user will have a very slow initial loading time.
The architecture needs to be updated such that:
A worker is responsible for updating the cache
Rather than using a memory cache, we setup an external redis server for caching. This has the added benefit that rather than each node in the cluster keeping its own cache, all nodes share one cache (the redis cache).
When the cache TTL expires, the cache will not be refreshed until the next user visits the site, because refreshing the cache is triggered by a request. This means that this user will have a very slow initial loading time.
The architecture needs to be updated such that: