Re: Which approach should we use for exposing metrics through Virtual tables?
I don’t think it’s really necessary. Or at least I’m not seeing why having individual, specialised virtual tables is not sufficient.
Nor do I think that we should expose everything that nodetool does, so IMO we shouldn’t aim for that. Even if the goal were to eventually deprecate and remove JMX/nodetool, we could do that without providing a CQL equivalent for everything it did.
On 22 June 2018 at 16:11:43, Chris Lohfink (clohfink@xxxxxxxxx) wrote:
However it is kinda necessary at some level for this "generalized" metrics.