Allow retrieving heatmap from spans/individual profiles #4736
Conversation
simonswine
force-pushed
the
20260105_add-heatmap
branch
from
ba47839 to
eb96f68
Compare
Label intersection when reusing labelSets was not run, also empty label set used.
simonswine
requested review from
a team,
aleks-p,
bryanhuhta and
marcsanmi
as code owners
simonswine
changed the title
pkg/model/heatmap/range.go
Outdated
| } | ||
|
|
||
| // createYAxisBuckets creates evenly spaced Y-axis buckets | ||
| func createYAxisBuckets(minValue, maxValue uint64, numBuckets int) []yBucket { |
There was a problem hiding this comment.
Can we use []int64 for Y buckets like for time buckets or does tracking min and max make some things easier?
There was a problem hiding this comment.
The mistake was made in the Exemplar proto API, not too sure this should either be int64 or straight a double: int64 is used in parquet and that is using it because of pprof. Our API Repsonse is using doubles (so is SelectSeries.
I would argue those values should be an int64. Will provide a PR in main after this merges
There was a problem hiding this comment.
Will provide a PR in main first, before fixing it here
pkg/model/heatmap/range.go
Outdated
|
|
||
| // findTimeBucket finds the index of the time bucket for a given timestamp | ||
| func findTimeBucket(timestamp int64, buckets []int64) int { | ||
| for i, bucket := range buckets { |
There was a problem hiding this comment.
Would a binary search be faster for the already sorted buckets?
There was a problem hiding this comment.
Depends on the amount of points, it should really be calcuated
| } | ||
|
|
||
| // This attribute table is used to store the attribute values for the heatmap. It might be reused for other queries at a later time. | ||
| type attributeTable struct { |
There was a problem hiding this comment.
I created a specific attributetable model in my PR. If it makes sense, we might want to reuse it once merged.
There was a problem hiding this comment.
Yes let's do that whatever merges first
3 participants
This PR implements a complete heatmap query API for visualizing profile data distribution over time and value dimensions, with support for both individual profiles and span-level data.
Key Changes: