Aggregate then Downsample

[Nicholas Ross]

I've been researching how to extract data with a different order of operations to the standard. I want to aggregate my data before down-sampling. Currently, post-analysis of the resulting data is not consistent when viewed at different granularities, and the order of operations is the cause of the issue - At one granularity I might get some data from Set A, but at another its all Set B. I need to merge the sets first so the data I'm down-sampling is consistent.

I found this old 2011 post on this board indicating that work may have been done toward this goal:

https://groups.google.com/forum/#!msg/opentsdb/r5O-BdYP9LA/-K5tJp90jCsJ

However I can't see any evidence of it. Is it possible to extract data in the way that I want?

I am using v2.3, but would be willing to upgrade versions if necessary.

Cheers, Nick

[ManOLamancha]

Hiya, not *quite* yet because the 2.x branch operations are fixed. 3.0's pipeline is now a graph so we can shuffle operations around. I'm almost done with the 2.x query format, then I need to get expressions working then I want to fix up the graphite style branch to include the downsampler so you can order ops as you like.

On top of that we'll likely implement PromQL on TSD and perhaps a custom DSL.