I am unsure what the best practices is for my use case:
We running a web-service and now want to achieve some redundancy. So we added a second server but with limited functionality.
The idea is to run the load on the second server all the time. Only if the first server goes down the traffic should swap to the “emergency” server with limited functionality.
As fare as I can tell I could do this on two ways with the Cloudflare loadbalancer.
Setup a single load balancer pool with both server. But the first server has the weight: 1 and the second server the weight: 0. So 100% load should be applied to the first server. And should swap to the second server in a event of failure on the first server.
Setup two load balancer pools. The first pool only with the first server. The second pool only with the second server. Both with the weight: 1. Then select the first pool as origin pool. And the second pool as fallback pool.
What would be the most practical thing to do here?