Historical documentary on Prometheus: open source telemetry for complex software systems(youtube.com)

Scientific instruments typically use proprietary, closed data structures and protocols, leading to vendor lock-in and a fragmented tooling ecosystem. In software, on the other hand, there is a mature technology stack for collecting telemetry data from complex systems of servers/processes (Prometheus [1], Grafana, Datadog, etc). This ecosystem is enabled by the OpenTelemetry protocol [2]. Since it's an open protocol (publicly documented, permissively licensed, free to use, etc), all telemetry tools implement it and gain from the inter-operability in a positive-sum game. I suspect that solving the telemetry problem would be a massive boost for reproducibility and AI-for-Science, so we would do well to learn from how telemetry is done for complex systems of software. [1] https://prometheus.io/ [2] https://opentelemetry.io/