Approval Workflows and Long-Running Processes Enterprise systems contain processes that live longer than one HTTP request. Model the workflow explicitly instead of scattering booleans such as approved=true across controllers.
- - - - DRAFT SUBMITTED REVIEW APPROVED EXECUTED > > > > Figure 3 - A workflow is a state machine with actors, transitions, and evidence
Workflow design checklist • Define states, allowed transitions, actor/permission requirements, transition preconditions, timestamps, reason/comments, and terminal states.
• Put transition rules in one authoritative backend policy/service; the UI only presents actions that may be available.
• Use explicit actions such as submit(), approve(), reject(), cancel(), finalize() instead of generic update({status}).
Long-running workflow • An approval may wait hours or days. Persist workflow state; do not keep it in memory or an HTTP connection.
• For multi-step asynchronous work, store progress and make each step restartable. Background workers can continue after process restarts. approvePurchaseOrder(poId, actor) { assertPermission(actor, "po.approve") assertState(po, "SUBMITTED") assertApprovalThreshold(actor, po.total) return transition(po, "APPROVED") }
An approval is a business transition, not merely a status field edit. Source basis: application state-machine/workflow design; persisted state is required for long-running business processes.
CI turns repository rules into automated gates. A useful pipeline must be understandable, reproducible locally, secure, and able to fail for the right reasons.
Restate the goal of “Approval Workflows and Long Running Processes” as one observable behavior. Write the success signal before touching code.
Trace the boundary map: git event → runner → install → quality gates → artifact/deploy decision. Mark the one boundary this lesson changes.
Inspect the current repository state and record the exact files, route, table, or runtime involved. Avoid creating an abstraction before you can name the existing owner.
Implement the smallest vertical slice. Add one coherent change, run the narrowest verification command, and read the output before continuing.
Connect the slice to TaskFlow’s phase artifact: Enterprise expansion blueprint for an HRIS/procurement/inventory/payroll-style system with modules, workflows, events, audit, and scaling triggers.. Verify behavior from the outside as a user or API client would.
Create one controlled failure that is safe to reproduce. Predict the symptom first, capture evidence, locate the earliest broken boundary, then fix only that cause.
Close the reference and rebuild or explain the core change from memory. Record what you would monitor or test before trusting the same change in production.
name: CI
on: [pull_request]
jobs:
verify:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- run: corepack enable
- run: pnpm install --frozen-lockfile
- run: pnpm lint && pnpm typecheck && pnpm test && pnpm buildDo not copy this as a finished answer. Predict what each boundary does, type the smallest useful slice, run it, then compare the result with your prediction.
Compare Node/pnpm versions, lockfile state, environment variables, OS assumptions, and generated artifacts.
Set least-privilege permissions at workflow/job level and isolate untrusted pull-request code from secrets.
Without copying the final source, create a second small scenario that applies “Approval Workflows and Long Running Processes” to a different TaskFlow resource or input. Write the acceptance criteria first, implement it, and provide the evidence that each criterion is satisfied.
Model one long-running approval as a state machine with actors, transitions, evidence, and audit records.
Use the smallest change that demonstrates this lesson. Stop after it works and capture evidence.Dual writes to database and message broker without reliability design
Predict the symptom before causing the failure. Then locate the earliest boundary where reality diverges from expectation.Why does this responsibility live here, what input does it trust, what output proves success, and what would go wrong if another layer owned it?
If your explanation is vague, repeat the boundary trace before moving on.Close the reference. Reproduce the core idea from memory or explain the exact request/data flow aloud. Then check your diff or notes for accidental shortcuts that bypass validation, authorization, error handling, or verification earned in earlier phases.