test(perf): increase Trino worker execution concurrency

[bill-ph]

Summary

Follow up on the completed experiment in #1162, closed without merging its shape-selection code. Start from main and change only the isolated Trino test lane's worker execution concurrency:

• Set task.max-worker-threads=8 and task.min-drivers=16 per worker, 4x the pinned engine's expected one-CPU defaults of 2 and 4.
• Keep three 1-CPU/4-GiB workers, 3G heaps, query-memory limits, coordinator configuration, and aggregation task.concurrency unchanged.
• Preserve the existing SQL, cache settings, warmups, and measured iterations. No matrix, new profiles, or workflow changes.
• Document scope, benchmark dispatch, tradeoffs, and recovery. No test changes.

The shared isolated manifest is used by both Trino E2E and frozen perf. Shared and production deployments are unaffected.

Motivation

The four-shape benchmark showed little benefit from consolidation at equal resources and approximately 2x throughput at twice the worker budget. Historical CPU usage during exact distinct was around 48-56% of quota. The experiment increased both resources and CPU-derived concurrency, so it did not isolate a CPU or storage bottleneck.

This change tests whether additional overlapping scan work improves latency at the original resource budget. task.min-drivers is a leaf-driver target, not a hard limit or an S3 request count. Higher concurrency can also increase heap pressure and context switching; no performance improvement is claimed before validation.

Validation

• go test -count=1 ./tests/mw-dev ./tests/perf/publishercli passed.
• just test-scenario passed.
• git diff --check passed; independent review found no actionable issues.
• just lint reports six existing SA4023 diagnostics in unchanged control-plane code; reproduced on base d7e852d8. Changed-code lint (golangci-lint run --new-from-rev=origin/main) passed.

Run the existing benchmark on this branch and compare query medians, errors, memory pressure, and CPU behavior against the recorded baseline before merging:

gh workflow run scenario-dev.yml --ref codex/trino-perf-read-concurrency -f scenario=posthog_frozen_perf

[github-actions]

Test Impact Plan

Deterministic summary of how this PR changes tests, CI runners, and coverage-risk signals.

Summary

Area	Added	Changed	Deleted
Test files	0	2	0
E2E/journey files	0	0	0
Workflow files	0	0	0

Signals

• Test cases: +0 / -0
• Assertions: +0 / -0
• Skips or known failures added: 0
• Workflow continue-on-error added: 0
• Workflow path filters added: 0
• Test commands removed from justfile: 0
• E2E/journey retry lines added: 0

Coverage risk: neutral or increased

No coverage-reduction warnings detected.