If you do I strongly recommend storing the indices with the chunks[0] and configuring a hassle-free, scalable storage solution [1] up front. Migrating chunk storage after the fact is a royal pain. Do also note that the official Helm chart will only get you so far before you have to start horizontally scaling individual components [2]; depending upon your query and log volume you might want to scale these separately and the chart [3] doesn't support this. This GitHub issue [4] has some Kubernetes YAML examples if you need them.
