ref(span-buffer): Skip the EXISTS check by using child_ic heuristic#117804
Merged
cmanallen merged 1 commit intoJun 22, 2026
Conversation
github-actions
Bot
added
the
Scope: Backend
untitaker
approved these changes
Jun 17, 2026
untitaker
approved these changes
Jun 17, 2026
cmanallen
added a commit
that referenced
this pull request
Jun 22, 2026
…#118152) The only item added to the queue is a segment_key. ```python zadd_items = queue_adds.setdefault(queue_key, {}) zadd_items[result.segment_key] = new_deadline ``` A span can be in the queue if that span was once a segment root. A span that always knew its parent never becomes a segment_key so zrem-ing it does nothing. Given our max (and near average) batch size is 2000 we're zrem-ing <2000 keys that have never existed in the queue. In `add-buffer.lua` we know who the former segment roots were. From #117804: "Presence of child_ic implies that this span is a root span". Let's return those merged root spans and remove them from the queue. Updating the queue takes 17ms (9us per span) on average. Removing 2000 no-ops will probably have a large impact. In local profiling it was about a microsecond per span (2ms per batch). In production this number could vary.
sehr-m
pushed a commit
that referenced
this pull request
Jun 23, 2026
…117804) Existence of the `child_members_key` is implied by the existing of the `child_ic` value. We can skip the O(spans) `EXISTS` check and use the cached value instead.
sehr-m
pushed a commit
that referenced
this pull request
Jun 23, 2026
…#118152) The only item added to the queue is a segment_key. ```python zadd_items = queue_adds.setdefault(queue_key, {}) zadd_items[result.segment_key] = new_deadline ``` A span can be in the queue if that span was once a segment root. A span that always knew its parent never becomes a segment_key so zrem-ing it does nothing. Given our max (and near average) batch size is 2000 we're zrem-ing <2000 keys that have never existed in the queue. In `add-buffer.lua` we know who the former segment roots were. From #117804: "Presence of child_ic implies that this span is a root span". Let's return those merged root spans and remove them from the queue. Updating the queue takes 17ms (9us per span) on average. Removing 2000 no-ops will probably have a large impact. In local profiling it was about a microsecond per span (2ms per batch). In production this number could vary.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Existence of the
child_members_keyis implied by the existing of thechild_icvalue. We can skip the O(spans)EXISTScheck and use the cached value instead.