Integrations need accountable owners after launch

After an integration launches, it should remain clear who decides on its business effects and operation. When a change or incident occurs, a responsible person must be reachable and able to decide.

Messages can keep flowing while their meaning changes

In a fictional company, the ERP changes the meaning of a status notification. The connection still carries messages and technical reachability shows no fault. In the portal, however, the notification still means “Order confirmed”, although it now denotes receipt alone. The problem lies in how the transferred content is interpreted.

Someone needs to recognise that this change affects the portal workflow. It must also be clear who assesses the new meaning and approves the adjustment. The original setup does not answer those questions indefinitely.

Responsibility must enable action

We recommend distinguishing responsibility for business meaning from technical responsibility. The person responsible for business meaning establishes what the portal may tell users. Technical responsibility covers monitoring, troubleshooting and implementing agreed changes. In a small team, the same people can hold both responsibilities.

The Google SRE Workbook describes clear roles for incident coordination, operational response and communication. This model shows one possible allocation. It does not require a large on-call team for a small interface.

The scope depends on the effect

A rarely used internal display may need different response rules from a portal that accepts orders every day. Specify when the impact calls for interruption or escalation and who provides cover during an absence. A public response-time promise should reflect only what the organisation can support.

You can check responsibility with a concrete case: the connection is reachable, but the order status is unclear. If the team can identify who clarifies the statement and who controls further processing until then, the assigned role becomes a usable operating rule.

Sources

• Google SRE Workbook: Incident Response