1. 1

Based on this analysis, it should be possible, in theory, to develop an über-system. The write path could accept raw events, de-mux them into metrics, traces, and logs heuristically or by the “shape” of the data, make the relevant optimizations, and store them into purpose-driven backends. The read path could provide a single pane-of-glass-style interface abstracting over those backends.

  1.