docs: ADR 0005 competency mastery storage, without history

[jesperhodge]

Draft. Replaces the ADR 0005 in #657, which decided storage and history together.

Stacked on #713 (ADR 0004), so the diff here is only what ADR 0005 adds. Retarget to main once #713 merges.

What changed

The whole history and audit-trail question moves out to a future ADR. What is left is the MVP storage decision:

1. Materialize a status row at all three levels including the leaf: at most one row per learner and node, updated in place, absent until attempted.
2. That write is an upsert against the unique index, since two workers handling a redelivered event can both find no row and both insert.
3. All three tables in the default database, unpartitioned, read and written on the primary. This is what makes ADR 0004's single transaction and group row lock work.
4. No schema conventions of their own: BigAutoField primary key, real FK to settings.AUTH_USER_MODEL.

Plus an Assumptions section (the deployment routes the grade write to the same alias; nothing redirects the recorder's reads to a replica) and an Out of Scope section that records the audit requirement, what the MVP can and cannot answer, and the constraints the future history ADR inherits.

All seven rejected alternatives survive, reworked. Several of their cons cited the HISTORY table and had to be re-derived; most came back stronger against ADR 0004 than they were against history.

Changes to ADRs 0002 and 0003

Two of these are forward-compatibility, deliberately taken now because both are free today and a migration on a billion-row table later:

• CompetencyMasteryStatuses.rank. The seeded statuses are listed highest-first and there is no ordering column, so ADR 0004 Decision 2's "store the higher of the two" cannot be a single UPDATE. Doing it in Python would need a leaf row lock that ADR 0004's rejected alternative 1 explicitly claims the design does not need.
• modified on the three learner status tables, with created narrowed to the row's first write. Under in-place updates created no longer dates the current status. Without this, no MVP-era row can be dated, which makes a later history table's backfill meaningless.
• CompetencyRuleProfile.archived said archived profiles keep learner status history resolvable. It keeps rows resolvable.
• ADR 0003's Context sized learner data as learners x criteria x time. The time factor was append-only residue.
• ADR 0003 Decision 5's "how history is retained is not decided here" now points at where it is deferred to.

ADR 0001 and ADR 0004 need no changes.

Open questions for review

• Staff-facing reads have no index. Every learner status index leads with the learner, so "every learner's status for competency X" is exactly the non-indexed query the ADR's own Context calls painful at this row count. Named in Out of Scope rather than decided.
• User deletion / retirement. Decision 4 keeps a real FK to the user model but does not say what happens on delete. Named in Out of Scope.
• Native MySQL partitioning is foreclosed, not merely deferred: MySQL needs every partitioning column in every unique key, and the surrogate id and the unique (learner, node) key share none. Stated and accepted rather than glossed as reversible.

Verification

sphinx-build over docs/ succeeds with zero warnings. doc8 is clean except one pre-existing trailing-whitespace error on main, untouched here.

🤖 Generated with Claude Code

[openedx-webhooks]

Thanks for the pull request, @jesperhodge!

This repository is currently maintained by @axim-engineering.

Once you've gone through the following steps feel free to tag them in a comment and let them know that your changes are ready for engineering review.

🔘 Get product approval

If you haven't already, check this list to see if your contribution needs to go through the product review process.

• If it does, you'll need to submit a product proposal for your contribution, and have it reviewed by the Product Working Group.
  • This process (including the steps you'll need to take) is documented here.
• If it doesn't, simply proceed with the next step.

🔘 Provide context

To help your reviewers and other members of the community understand the purpose and larger context of your changes, feel free to add as much of the following information to the PR description as you can:

• Dependencies

  This PR must be merged before / after / at the same time as ...

• Blockers

  This PR is waiting for OEP-1234 to be accepted.

• Timeline information

  This PR must be merged by XX date because ...

• Partner information

  This is for a course on edx.org.

• Supporting documentation
• Relevant Open edX discussion forum threads

🔘 Get a green build

If one or more checks are failing, continue working on your changes until this is no longer the case and your build turns green.

🔘 Update the status of your PR

Your PR is currently marked as a draft. After completing the steps above, update its status by clicking "Ready for Review", or removing "WIP" from the title, as appropriate.

---

Where can I find more information?

If you'd like to get more details on all aspects of the review process for open source pull requests (OSPRs), check out the following resources:

• Overview of Review Process for Community Contributions
• Pull Request Status Guide
• Making changes to your pull request

When can I expect my changes to be merged?

Our goal is to get community contributions seen and reviewed as efficiently as possible.

However, the amount of time that it takes to review and merge a PR can vary significantly based on factors such as:

• The size and impact of the changes that it introduces
• The need for product review
• Maintenance status of the parent repository

💡 As a result it may take up to several weeks or months to complete a review and merge your PR.