docs: refresh PM report for final submission state

[0b00101111]

Summary

Bring the project-management report forward to 2026-04-20 for the final submission. The previous version (PR #168) froze at Apr 18 and predates the final sprint — experiments, deck iterations, rename, lessons cleanup.

Changes

• Header counts refreshed: 292 commits · 127 merged PRs · 157 passing + 32 skipped tests
• PR split: Sam 77 (61%) · Jess 50 (39%)
• Timeline: new Apr 18–20 row covering the final sprint
• Problems table: three new entries (11–13) — straggler-chart concurrency assumption, Campaign→Script naming, lessons-series V1/V2 cleanup
• Process discipline: added 'experiments drove design changes' bullet tying straggler experiment → PR feat: transition at 95% completion to skip stragglers #165, Locust → CPU-signal mistake, fan-out → Amdahl's knee at cap=29
• Final State section now lists the experiment portfolio and full deliverables

Notes for reviewer

• Numbers verified from git rev-list --count main, gh pr list, and pytest --collect-only.
• This is a team deliverable for the CS6650 final submission (5 marks).

Test plan

• Read the final doc end-to-end; check that timeline, splits, and problems list match your recollection
• Call out anything in the new Apr 18–20 row or problems 11–13 that misrepresents the work

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Updates the CS6650 project-management report to reflect the project’s final submission state as of 2026-04-20, including refreshed metrics and additional final-sprint items.

Changes:

• Refreshes header metrics (duration, commit/PR/test counts) and PR split.
• Extends the work breakdown timeline with a new Apr 18–20 final-sprint row.
• Adds new “Problems Encountered” items (11–13) and expands “Final State” details (experiments + deliverables).

[Copilot]

The duration math looks off: 2026-03-22 → 2026-04-20 is ~29 days (~4.1 weeks), not ~4.5 weeks. Suggest adjusting the approximation (or dropping it) so the header stays internally consistent.

Suggested change

	**Duration:** 2026-03-22 → 2026-04-20 (~4.5 weeks)
	**Duration:** 2026-03-22 → 2026-04-20 (~4.1 weeks)

[Copilot]

These bullets still say "112 PRs" / "112 merged PRs", but the header scope was updated to "127 merged PRs". Please update these counts (or rephrase to avoid hard-coding) so the report doesn’t contradict itself.

Suggested change

	- **One PR, one fix** — 112 PRs averaged 4 files each. Every PR title used conventional prefixes (`feat:`, `fix:`, `chore:`, `docs:`) for grep-friendly history.
	- **Hard branch protection on `main`** — Sam and Jess set a GitHub ruleset that *physically blocks* direct pushes and requires ≥1 approving review from the other team member before merge (admin override disabled). This turned "we should review each other's work" from a norm into an enforced gate. Effect: every one of the 112 merged PRs has at least one reviewer's signature; neither of us can ship without the other reading the diff. On multiple occasions Jess's review caught issues Sam missed (and vice-versa) before they hit production.
	- **One PR, one fix** — 127 PRs averaged 4 files each. Every PR title used conventional prefixes (`feat:`, `fix:`, `chore:`, `docs:`) for grep-friendly history.
	- **Hard branch protection on `main`** — Sam and Jess set a GitHub ruleset that *physically blocks* direct pushes and requires ≥1 approving review from the other team member before merge (admin override disabled). This turned "we should review each other's work" from a norm into an enforced gate. Effect: every one of the 127 merged PRs has at least one reviewer's signature; neither of us can ship without the other reading the diff. On multiple occasions Jess's review caught issues Sam missed (and vice-versa) before they hit production.

[Copilot]

The parenthetical says the 32 skipped integration tests are "gated on AWS creds", but the skipped experiment/integration modules are actually gated on environment variables like PIPELINE_BASE_URL / ELASTICACHE_URL being set (see backend/tests/experiments/test_e2e_pipeline.py and test_elasticache_integration.py). Suggest updating the wording to match the actual skip conditions.

Suggested change

	- **Tests before features** — 105 unit tests existed before the first real pipeline run. Grew to 157 passing (+ 32 skipped integration tests gated on AWS creds). Caught the `RateLimitError` retry countdown bug pre-deploy.
	- **Tests before features** — 105 unit tests existed before the first real pipeline run. Grew to 157 passing (+ 32 skipped integration tests gated on integration env vars such as `PIPELINE_BASE_URL` / `ELASTICACHE_URL`). Caught the `RateLimitError` retry countdown bug pre-deploy.