Revision history [back]

Diagnostics were really originally designed for human readable outputs. The diagnostics_aggregator really exists to clean up diagnostics and re-arrange them for human viewing - the aggregator uses Analyzers (there is a tutorial on creating custom analyzers).

The messages are generally not ideal for machine processing (while there are warning levels, most of the information tends to be string-based) - but you could certainly do that. You would almost certainly have to write your own node which could subscribe to either of the aggregated or non-aggregated diagnostics topics.