I like the idea, especially that it's written in C (the core part, anyway) and there's just a few dependencies to install. However, the part of README that show

I like the idea, especially that it's written in C (the core part, anyway) and there's just a few dependencies to install. However, the part of README that shows how to run it is completely unreadable for me, as it uses magic Docker incantations that I can't separate from SiriDB's own command line options. Also, a collection of time series is not a flat space, contrary to your data model, where a time series only has a name. Usually there are multiple dimensions, e.g. (location) x (server name) x (cpu number) for CPU usage collected from servers from multiple server rooms. Note that it's not hierarchical and the dimensions can differ between metrics (mount point for disk space, device name for I/O operations, <nothing> for RAM usage, and so on).