Execute Planner residual DAGs and record reproducible o11y comparisons - #530
Merged
Merged
Conversation
# Conflicts: # control_plane/src/physical/compiler.rs # control_plane/src/physical/workload_cost.rs # control_plane/src/query_plan.rs # data_plane/src/drivers/ingest/prometheus_remote_write.rs # data_plane/src/precompute_engine/worker.rs # tools/o11y-execution/replay.py # tools/o11y-execution/test_replay.py
…y-acceptance # Conflicts: # control_plane/src/physical/compiler.rs # control_plane/src/physical/workload_cost.rs
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Before this PR, retained Planner logical subtrees generally became whole-query external fallback, so binding coverage could not establish executable workload coverage. After this PR, the control plane compiles the selected residual tree into a typed installed DAG, preserving safe summary siblings and executing raw branches with explicit source, labels and evaluation times.
Validation: real upstream o11ybench corpus contains 28 occurrences / 24 unique queries. Independent local operator comparison passed 112 evaluations over 4 timestamps. Frozen pre-integration benchmark ran 4 complete matched trials: 2240/2240 equal, all external exact fallback, zero warm/hybrid. Backend+fallback used 4.90–5.26 query CPU seconds per trial versus Prometheus 2.41–2.74 seconds. This is explicitly a no-benefit result, not completed acceleration acceptance. Frozen binaries, source/data hashes, candidate/selection evidence and per-query responses are archived in /mydata/o11y-acceptance-results.
Limitations: finite-input drain seals ingestion; repeated queries retain original timestamps. Fresh processes do not establish cold OS caches or concurrent throughput. Full lifecycle benefit and general PromQL coverage are not claimed. New filtered range-max materialization is separate follow-up work.
Final integrated validation: 682 control-plane library tests, 110 execution-engine tests, 5 real compatibility process tests and 31 Python tests passed. The full data-plane library suite passed 1001 tests before the final pane integration; focused engine/process checks cover that integration.