So, the server that runs the website also fetches the data on demand, which it then caches for up to 1 minute in order to not get denied with 429 Too Many Requests. That was the optimal balance between freshness and stability in my tests.
I'm using a cheap server provider for now, so my server costs amount to around $5/month. I expect that to (hopefully) scale up, but for now it works without issues.
So, the server that runs the website also fetches the data on demand, which it then caches for up to 1 minute in order to not get denied with 429 Too Many Requests. That was the optimal balance between freshness and stability in my tests.
I'm using a cheap server provider for now, so my server costs amount to around $5/month. I expect that to (hopefully) scale up, but for now it works without issues.