QORPO Blog / Industry

Unreal network dormancy: checking an interactable before and after it changes

Q
QORPO Studio20 November 2026 · 4 min read
Two miniature metal doors in closed and opening states on a dark testing table
Concept illustration created for this article, not a gameplay screenshot.

When reviewing network dormancy in Unreal, test an interactable’s state before a change, immediately after the change and when a new client observes it. A quiet object that looks correct on the server is not enough evidence that every relevant client receives its current state.

Dormancy concerns replication work for an actor that changes infrequently. Relevancy concerns whether that actor matters to a particular connection. Keeping those concepts separate makes a small multiplayer review easier to interpret.

Choose one simple state change

Use an isolated model object, such as a replicated door with open and closed states. Define the expected state transition and the clients that should observe it.

Record the engine version, network setup and the actor configuration. Keep animation, sound and unrelated interaction effects out of the first diagnostic case where possible.

This is a proposed review method, not a claim that QORPO shipped or tested this particular door. It provides a small example for examining replication behaviour without diagnosing an entire multiplayer level at once.

Read the dormancy contract first

Epic’s Actor Network Dormancy documentation explains that dormant actors are excluded from the usual replication consideration until their dormancy is handled appropriately.

The same documentation describes waking or flushing dormancy before changing replicated properties. Inspect how the project handles that transition rather than assuming that changing a value automatically reaches existing clients.

Do not copy a setting merely because the model object is a door. The useful question is whether the actual actor changes infrequently and whether its change path follows the intended replication contract.

Observe two clients separately

Place two clients where the object should be relevant. Begin with a known closed state, perform the supported interaction and compare both client views with the authoritative state.

Record which view changed, which remained stale and when the transition occurred. A server-only screenshot cannot establish the joining or observing player’s experience.

If both clients update, repeat the ordinary second interaction. A correct first change does not prove that the actor returns to and wakes from its intended dormant state correctly later.

Add a new observer

A client that already saw the object and a client observing it for the first time are different cases. Include a supported join or arrival sequence after the state has changed.

For the hypothetical door, the new observer should receive the state appropriate to that moment. It should not rely on having witnessed an earlier opening animation.

Describe the actual result without guessing the cause. A late observer seeing a closed door can have several explanations, so the report needs the starting conditions and state rather than an unsupported diagnosis.

Keep relevancy separate

Epic’s Actor Relevancy guide explains that relevance is determined for each connection. It is not identical to whether an actor is dormant.

If one client cannot see or receive the actor because of the project’s relevance rules, that is a different observation from a visible actor retaining an old property value.

Use the project’s actual relevance setup when choosing viewpoints. Do not force every actor to remain relevant simply to make a narrow test appear successful. That would change the question and potentially the intended networking workload.

A model failure and a useful report

Imagine a hypothetical interaction where the server door opens, one existing client remains closed and a newly joining client sees the open state.

That combination is useful evidence because it separates an existing observer from a new one. It does not, on its own, prove exactly which call or property caused the difference.

Save the actor’s initial state, the interaction, each client’s view and the order of joining. A concise reproduction gives an engineer a much better starting point than “multiplayer doors are unreliable.”

Measure the intended benefit separately

Correct state delivery is the first requirement. If dormancy is introduced to reduce server work, measure that work under a representative load after the behaviour checks pass.

Do not turn Epic’s general discussion of possible savings into a measured performance claim for your project. A small fixture can establish behaviour while being too small to represent the real workload.

Our repeatable Unreal Insights route guide describes recording a consistent sequence. The same discipline helps when comparing a candidate configuration with a baseline.

Extend the conditions deliberately

Once the basic case works, add the network conditions and object counts the product actually supports. The network-emulation guide covers that separate testing step.

Keep track of what changed between runs. A behaviour that passes on a local connection still needs evidence under the intended conditions, but a failure under a different setup should not erase the original observation.

Save the build, configuration, executed cases, measurements and remaining coverage. Use the test-fixture guide to make expected states explicit.

Follow QORPO’s live development for more production-focused articles. A successful dormancy change preserves the player’s current view of the world while making its technical benefit measurable.

Sources