The next logical step would be to configure Prometheus with additional targets. Specifically, you may want to configure it to fetch metrics that are already made available through the Kubernetes API. We, however, will NOT be doing this. First of all, this chapter is not about monitoring and alerting. The second, and the more important reason, is that using the hostPath Volume type to provide configuration is NOT a good idea.

A hostPath Volume maps a directory from a host to where the Pod is running. Using it to “inject” configuration files into containers would mean that we’d have to make sure that the file is present on every node of the cluster.

Working with a single-node cluster can be potentially misleading. The fact that we’re running a single-node cluster means that every Pod we run will be scheduled on one node. Copying a configuration file to that single node, as we did in our example, ensures that it can be mounted in any Pod. However, the moment we add more nodes to the cluster, we’d experience side effects. We’d need to make sure that each node in our cluster has the same file we wish to mount, as we would not be able to predict where individual Pods would be scheduled. This would introduce far too much unnecessary work and added complexity.