Current decision

84% klar, provider credentials och webhook proof kvar

Flow har integrationsramverket och proof-ytorna. 100% kräver riktiga tenant/provider credentials, webhook replay-evidence och reconciliation mot live-data.

4

proof lanes

6

checks

4

provider proof

Connector proof lanes

Det som gör integrationerna pålitliga i kundens ögon

Provider setup and diagnostics

Each connector family has setup, health, provider status and owner visibility.

  • setup state
  • credential gate
  • health status
  • owner

Webhook replay and retry

Replay/retry paths are defined and visible. Full proof requires real provider webhook evidence.

  • webhook received
  • retry queue
  • manual replay
  • idempotency

Reconciliation

Flow can show expected vs delivered records and explain drift without exposing raw secrets.

  • external ID
  • internal ID
  • drift reason
  • repair action

SLA and degradation

Provider outage/degradation has customer-safe copy, owner and recovery expectation.

  • degraded provider
  • throttling
  • recovery ETA
  • customer-safe status

Acceptance

Go-live checks

  • No connector is called production-proof without tenant-owned credentials.
  • Every first-class connector has setup, health, retry and reconciliation visibility.
  • Webhook replay proof is archived for at least one success and one recovery path.
  • Provider degradation has owner, customer-safe copy and recovery ETA.
  • External IDs and internal Flow IDs can be reconciled without leaking secrets.
  • Connector proof separates framework readiness from certified provider evidence.

Remaining external proof

Det som kräver riktiga providers

  • Archive provider-owned credentials proof for HubSpot, Zendesk/Freshdesk and first PMS lane.
  • Run webhook replay proof per first-class connector.
  • Capture throttling/degraded-provider evidence from a real or approved sandbox provider.
  • Run reconciliation proof against live tenant records.