Concept Design

SaaS & technology

For product and platform teams operating enterprise SaaS across identity providers, billing platforms and isolated customer environments.

Recovering subscription access after delayed or duplicate billing events

A billing service can record a commercial change before the application applies it. Renewals, plan changes and cancellations may reach the product out of order. Retrying the last webhook can restore the wrong feature set.

Workflow at a glance
  1. Source recordsBilling events and contract mapping
  2. ProcessingResolve current entitlements
  3. Decision ownerProduct operations
  4. Operational outputVerified tenant access

Implementation approach

Store each verified event once, retrieve the authoritative entitlement state and map the billing customer to the correct tenant and contract. Apply versioned access changes and record acknowledgement from the product.

Exceptions and operating controls

Distinguish provider billing state from a contractual grace period. Make support overrides time-limited and attributable. Route unknown customer-to-tenant mappings to review without changing access.

Illustrative operating scenario

A cancellation event is retried after a later renewal. The consumer refreshes current entitlements before updating the tenant, then records the stale event as processed.

Acceptance test and handover

Test out-of-order delivery, duplicate events and provider timeouts against test accounts. Compare billed entitlement with effective product access and test replay from a checkpoint.

Industry evidence

Stripe exposes active entitlements so an application can determine which features to make available. Stripe Billing Entitlements

Stripe documents duplicate deliveries, retries and the absence of guaranteed event order. Stripe event delivery and ordering

Preventing unintended deprovisioning during an enterprise directory change

An enterprise customer can change a directory filter or group mapping and unexpectedly put a large set of users out of scope. Native provisioning safeguards help, but the product still needs to preserve tenant boundaries, delegated roles and a recoverable state.

Workflow at a glance
  1. Source recordsDirectory provisioning events
  2. ProcessingPreview tenant-specific impact
  3. Decision ownerCustomer administrator
  4. Operational outputRecoverable access change

Implementation approach

Integrate SCIM provisioning with the product’s tenant and role model. Preview scope changes, distinguish disablement from deletion, log affected accounts and provide a tenant-scoped recovery path for approved corrections.

Exceptions and operating controls

The customer administrator confirms intended scope. Quarantine ambiguous tenant mappings and review broad changes before release. Preserve audit history when a person is restored under the same identity.

Illustrative operating scenario

A customer renames its main access group. Provisioning pauses a batch of proposed removals. Support can inspect the tenant impact and verify the corrected mapping before resuming.

Acceptance test and handover

Test a legitimate leaver alongside a mistaken bulk change and a group rename. Check that restoration does not recreate privileged access the user no longer holds.

Industry evidence

Microsoft documents deletion thresholds, quarantine and staged actions in its provisioning service. Microsoft Entra provisioning safeguards

Moving a live customer between tenant environments

A database copy can succeed while scheduled jobs, files, webhooks or cached lookups still point to the old tenant location. The problem appears later when a background task writes to the previous environment.

Workflow at a glance
  1. Source recordsTenant resources and jobs
  2. ProcessingSynchronise and fence writes
  3. Decision ownerPlatform owner
  4. Operational outputValidated tenant cutover

Implementation approach

Inventory the tenant’s data, object storage, job queues, search indexes, keys and outbound integrations. Build a cutover sequence with a write barrier, incremental catch-up, verified routing change and a defined rollback point.

Exceptions and operating controls

The platform owner controls cutover. Maintain tenant isolation during validation and quarantine writes that arrive after the barrier. Identify the point beyond which rollback requires replaying new writes.

Illustrative operating scenario

An export job begins before migration and finishes afterwards. The cutover procedure detects its old destination and prevents duplicate delivery to the customer.

Acceptance test and handover

Migrate a representative test tenant under traffic. Validate object manifests, business totals, queued work and routing, then rehearse rollback with writes made after cutover.

Industry evidence

Microsoft describes tenancy and storage choices that affect isolation, scaling and movement of customer data. Microsoft Azure Architecture Center on multitenant data

Discuss a workflow All case studies