fix(core): Retry role mapping order races

[haimingZZ]

Summary

• retry role mapping rule creation when a concurrent create picks the same temporary (type, order) slot
• re-read the current ordering before retrying so the second create can append/splice after the rule that won the race
• add a backend concurrency-style regression test that starts two appends together and verifies both complete with contiguous ordering

Review notes

This builds on #28722. That change avoids ordinary requested-order collisions and serializes the UI, but two backend/API creates can still read the same snapshot and choose the same temporary order for the initial insert. This keeps the existing two-phase applyOrder flow and only retries unique insert conflicts after re-reading.

Validation

• git diff --check
• pnpm --filter n8n test:unit -- role-mapping-rule.service.ee.test.ts (not run: this worktree has no node_modules, so jest is not installed)

[cubic-dev-ai]

No issues found across 2 files

Architecture diagram

sequenceDiagram
    participant API as API/Provisioning Logic
    participant Service as RoleMappingRuleService
    participant Repo as RoleMappingRuleRepository
    participant DB as Database (Postgres/SQLite)

    Note over API,DB: Role Mapping Rule Creation Flow

    API->>Service: create(dto)
    Service->>Repo: find(existing rules for type)
    Repo-->>Service: rules[]
    
    Service->>Service: Calculate tempOrder (max + 1)
    
    rect rgb(240, 240, 240)
    Note over Service,DB: NEW: Retry Logic for Concurrent Inserts
    loop Until Success or Max Retries
        Service->>Repo: save(new rule with tempOrder)
        
        alt Success Path
            Repo->>DB: INSERT rule
            DB-->>Repo: OK
            Repo-->>Service: saved entity
        else CHANGED: Conflict (Race Condition)
            DB-->>Repo: Error: Unique Constraint (type, order)
            Repo-->>Service: throws ConflictError
            Service->>Repo: NEW: re-read current rules
            Repo-->>Service: latest rules[]
            Service->>Service: NEW: Recalculate tempOrder
        end
    end
    end

    Note over Service,DB: Two-Phase Ordering (applyOrder)
    
    Service->>DB: BEGIN TRANSACTION
    
    rect rgb(230, 245, 255)
    Note over Service,DB: Phase 1: Vacate slots (Move to high offset)
    loop for each ruleId in reorderedIds
        Service->>DB: UPDATE order = offset + i
    end
    end

    rect rgb(230, 255, 230)
    Note over Service,DB: Phase 2: Normalize (Sequence from 0)
    loop for each ruleId in reorderedIds
        Service->>DB: UPDATE order = i
    end
    end

    Service->>DB: COMMIT
    
    Service-->>API: RoleMappingRuleResponse

Loading

[n8n-assistant]

Hey @haimingZZ,

Thank you for your contribution. We appreciate the time and effort you’ve taken to submit this pull request.

Before we can proceed, please ensure the following:
• Tests are included for any new functionality, logic changes or bug fixes.
• The PR aligns with our contribution guidelines.

Regarding new nodes:
We no longer accept new nodes directly into the core codebase. Instead, we encourage contributors to follow our Community Node Submission Guide to publish nodes independently.

If your node integrates with an AI service that you own or represent, please email nodes@n8n.io and we will be happy to discuss the best approach.

About review timelines:
This PR has been added to our internal tracker as "GHC-8063". While we plan to review it, we are currently unable to provide an exact timeframe. Our goal is to begin reviews within a month, but this may change depending on team priorities. We will reach out when the review begins.

Thank you again for contributing to n8n.

[cla-bot]

The cla-bot has been summoned, and re-checked this pull request!