Capacity buffer
Also known as Capacity slack, Planning buffer
Definition
A capacity buffer is intentionally uncommitted capacity held available for variation or work that cannot be forecast precisely. It is a planning policy, not evidence that people should remain idle.
Slack absorbs variation
A delivery system with no available slack reacts to every surprise by creating a queue, interrupting planned work, or extending commitments. A capacity buffer gives the system room to absorb incidents, support requests, review demand, or uncertain work. The buffer should be visible so it is not mistaken for unused capacity that must be filled.
A concrete example
Suppose an illustrative team usually has ten delivery slots per month and reserves two for operational requests. If only one is needed, the remaining slot can be pulled into planned work late in the period. If five requests arrive, the team has a transparent basis for discussing the overflow.
Set a policy
Use historical variation and the cost of delay to choose an initial reserve. Review whether the buffer is consumed by recurring work that deserves its own service class. A buffer can also be expressed as a WIP ceiling or a percentage of planned capacity.
Limitations
There is no universal buffer percentage. An overly large reserve can hide weak prioritization, while a small one can make service brittle. Evaluate lead time, incidents, and missed commitments together.
How this relates to Weave
Weave can help teams examine historical delivery variation and unplanned engineering activity when setting a buffer policy. It does not determine the right reserve for a team, because operational obligations and business risk sit outside repository activity.
Explore Engineering intelligence