[milindsrivastava1997]

Summary

Adds 33 small, single-purpose workloads to sketch_plan_snapshots, each pinning one memo behavior through its selected plan. Closes #1278. Stacked on #1239 (refactor/1207-tools).

Review guide

• crates/devtools/tests/sketch_plan_snapshots/manifest.json: the 33 new entries. Each covers says what the workload pins and names the issue when the plan reflects a known gap.
• crates/devtools/tests/sketch_plan_snapshots.rs: one header paragraph.
• workloads/*.json and <id>.json snapshots are generated or copied; skim them.

area	ids
rollup across label sets	rollup-quantile-nested-labels, rollup-quantile-disjoint-labels, rollup-sum-chain, stddev-nested-labels, avg-nested-labels, sql-rollup-quantile-same-window, sql-rollup-quantile-different-windows
top-k	topk-5-and-10, topk-by-job-and-global, topk-5-and-10-accuracy, topk-rate-counter, topk-sum-count-over-time, topk-sum-count-over-time-accuracy
chains	quantile-over-time-no-phase, quantile-over-time-phase-0, chains-10m-5m-3m-3m, chains-10m-5m-3m-10s, chains-10m-5m-3m-10s-mvp, sum-over-time-two-intervals, quantile-phase-0-and-30s, quantile-7m-every-5m
over-time aggregates	sum-avg-over-time-5m, sum-5m-avg-10m-over-time, mixed-quantiles
sketch choice	distinct-over-time-1pct, distinct-over-time-5pct, entropy-over-time, quantile-two-targets, distinct-two-forms, entropy-and-distinct, quantile-exact, sign-unknown, sign-non-negative

All are on data (10k gauge series, 15 s samples), mostly every 1m at phase 0, under raw_retained (one under mvp). The two SQL workloads use workloads/sql-table-data.tables.json.

Behavior: before → after

rollup-quantile-nested-labels (quantile by (job) (0.9, data) + quantile by (job, host) (0.9, data)):

• before: no snapshot; the missing rollup (Rollup across label sets is never offered for PromQL queries (#1239) #1281) was visible only by running asap_planner by hand.
• after: the snapshot pins two Scans and two quantile plans. A fix to Rollup across label sets is never offered for PromQL queries (#1239) #1281 changes it to one Scan with SummaryMerge from the (job, host) KLL.

Interfaces

None.

Algorithm changes

None.

Decisions and open questions

• The snapshots record current behavior, including known gaps: asap_planner: Pass 2 builds no tumbling windows for fixed_interval without evaluation_phase #1260, Pass 2: tumbling-window chains aren't shared across queries with different grids #1270, Rollup across label sets is never offered for PromQL queries (#1239) #1281, Rollup across different windows through a shared chain is not offered (#1239) #1282. Fixing those is expected to change these snapshots.
• A snapshot pins only the Stage 3 plan. Gaps that don't change the selected plan (for example Pass 2: tumbling-window chains aren't shared across queries with different grids #1270 under the current cost model) are not caught; they need a Pass 2 assertion.
• Some workloads are exact. They pin sharing between queries, which the single-query corpus does not.

Verification

• UPDATE_SNAPSHOTS=1 cargo test --release -p asap-devtools --test corpus_snapshots -- --ignored sketch_plans_match_snapshots, then the same without UPDATE_SNAPSHOTS: passes (43 workloads, ~5.6 s).
• Snapshot changes: none of the 10 existing snapshots changed. The 33 new ones are added; they record the plans described in each covers.
• Reviewer confirms the recorded plans are the intended baseline: (for a person to fill in)

Follow-ups

• Rollup across label sets is never offered for PromQL queries (#1239) #1281, Rollup across different windows through a shared chain is not offered (#1239) #1282, Pass 2: tumbling-window chains aren't shared across queries with different grids #1270, asap_planner: Pass 2 builds no tumbling windows for fixed_interval without evaluation_phase #1260: their fixes update the matching snapshots.
• A Pass 2 candidate-level check for behaviors the plan does not show.

🤖 Generated with Claude Code

https://claude.ai/code/session_014qhwq4GZoWwtfFWWr9CKkr