This commit refactors the tsdb query engine to use separate aggregate
and raw execution paths, encapsulates cursor functionality, and removes
the TagSetCursor from the aggregate path. By removing the TagSetCursor,
we can pass sets of unordered values to the map functions and bypass
the `container/heap` entirely.
If an aggregate derivative query did not have a value in the first
time bucket, it would abort early and return a single row with value
of 0. Similarly, if either the current or previous value was nil,
it would skip the row and not append any values causing gaps and
no data to show up.
Instead, this will append a nil value if either the current or previous
valis is nil. This essentially allows nil values to carry through the
results as well as gives a more sensible value for rows where we cannot
compute a difference (instead of dropping the row as before).
Fixes#4237#4263
CPU profiling shows that computing the tagset-based key of each point is significant CPU cost. These keys actually don't change per cursor, so precompute once at mapper-open time, and then use those values as points are drained from the cursor.
Before this change the cursor tag was getting computed on every point, which involved marshalling tags.
Ben Johnson
Cory LaNou
Jason Wilder
Daniel Morsing
Jason Wilder
David Norton
Philip O'Toole
Philip O'Toole
Cory LaNou