Fix Godot NavigationAgent Stuck on Obstacles

Godot 4’s navigation system pairs a pathfinder on a baked mesh with an RVO (reciprocal velocity obstacle) avoidance step that steers the agent away from other agents and obstacles. Either half can fail independently. If your enemy stops a meter short of the player and shakes in place, the pathfinder has delivered a good path but the avoidance module is outputting a zero velocity because the scene is too crowded.

Check the agent radius against the passage width

The single setting that causes the most stuck-agent reports is radius. RVO refuses to emit a velocity that would bring the agent’s radius circle into contact with any obstacle or other agent. If the gap between two walls is 1.0 meter and your radius is 0.6, the clearance on each side is 0.4 — less than the radius — so RVO picks zero velocity as the safest choice.

var agent: NavigationAgent3D = $NavigationAgent3D
agent.radius = 0.35   # smaller than half the narrowest doorway
agent.height = 1.8
agent.max_speed = 4.0

Agent radius and navmesh agent radius are separate. The baked mesh is inset by whatever agent radius you used at bake time. The RVO radius is runtime only. Keep the RVO radius equal to or slightly smaller than the baked radius so the agent never wants to go where the mesh has no room.

Set path_desired_distance correctly

Every frame the agent looks at the next waypoint in its path and asks “am I close enough to move on?” That check uses path_desired_distance. If this value is smaller than one frame’s worth of movement at your max speed, the agent overshoots the waypoint, turns around, overshoots again, and never advances. A safe rule is:

agent.path_desired_distance = max(0.5, agent.max_speed * get_physics_process_delta_time() * 2.0)
agent.target_desired_distance = 1.0

target_desired_distance controls when the agent considers the final destination reached. Setting it to match your melee range or interaction radius stops enemies from trying to stand exactly on the player’s feet.

Decide which obstacles carve vs avoid

Godot distinguishes two obstacle behaviors. A NavigationObstacle with affect_navigation_mesh on is a carver: it modifies the baked mesh and forces the pathfinder to route around it. This costs a rebake whenever the obstacle moves. Without that flag, the obstacle is RVO-only — invisible to the pathfinder, visible to the avoidance step. Use carving for static or rarely-moving props (a closed door, a parked cart) and RVO-only for moving hazards (a patrolling enemy, a rolling barrel).

var obstacle: NavigationObstacle3D = $NavigationObstacle3D
obstacle.affect_navigation_mesh = true
obstacle.carve_navigation_mesh = true
obstacle.vertices = [...]

Watch for agent_max_speed throttling

RVO selects the best velocity whose magnitude is at most max_speed. If a script resets max_speed to zero during a stunned state and you forget to restore it, the agent visually freezes and the developer suspects a navigation bug. Log the velocity each frame to rule this out:

func _physics_process(delta):
    var next := agent.get_next_path_position()
    var dir := (next - global_position).normalized()
    agent.set_velocity(dir * agent.max_speed)
    print("vel", agent.max_speed, "next", next)

If next equals your current position, the pathfinder believes you are at the waypoint — back to the desired-distance check. If next is sensible but the agent will not move, confirm you are passing velocity through the velocity_computed signal, not applying the raw direction. In Godot 4 the RVO step returns its final velocity via that signal; the raw direction will fight avoidance.

Trigger targeted rebakes

For dynamic level elements that carve the mesh, call NavigationServer3D.map_force_update sparingly after they move. A full rebake is expensive; prefer to limit carving obstacles to a handful of objects. If you have more than a few dozen moving carvers, you are fighting the architecture — switch them to RVO-only and accept some path re-evaluation.

Understanding the issue

Navigation meshes are precomputed. Changes to geometry invalidate the precomputation; runtime regenerates may take long enough that the player notices. Plan invalidation timing carefully.

The specific bug described above is the kind that surfaces during integration rather than unit testing. It depends on a combination of factors: the asset configuration, the runtime state, the platform's specific behavior. In isolation, each piece looks correct; in combination, the bug emerges. This is why thorough integration testing - playing the actual game in realistic conditions - catches things that automated tests miss.

Why this happens

Bugs of this class are particularly easy to ship past internal QA because they often depend on specific runtime conditions - hardware combinations, network states, or asset configurations that QA didn't reproduce. Players hit them in the wild, file reports that are hard to repro, and the bug accumulates negative reviews while engineering tries to recreate the failure mode.

At the engine level, the behavior comes from a deliberate design decision in Godot. The engine team chose a particular trade-off - usually performance versus convenience, or generality versus specificity - and that trade-off has consequences when you push against it. Understanding the trade-off is what turns 'this bug is mysterious' into 'this bug is the expected consequence of this design'.

Verifying the fix

For shipping games, the safest verification is a staged rollout. Apply the fix to 1% of players for 24 hours; watch the affected metric; expand if green. Skipping the staged rollout means the verification is the entire player base, which is too high a stakes for most fixes.

Reproducibility is the prerequisite for verification. If you can't reliably reproduce the bug pre-fix, you can't reliably verify it post-fix. Spend time getting a clean reproduction before you write any fix code. The fix is fast once you understand the reproduction; the reproduction is the slow part.

Variations to watch for

Related bug classes often share the same root cause. If you find yourself fixing this issue, look for cousins: similar symptoms in adjacent systems, the same data flow but a different value, or the same fix pattern in another module. The catalog of 'we've seen this before' becomes valuable institutional knowledge.

Adjacent bugs often share a root cause. After fixing the case you've found, spend an hour searching the codebase for similar patterns. What's the same call with different arguments? The same data flow with a different entity type? The same lifecycle issue in a sibling system? Each match is a candidate for the same fix, or a related fix that prevents future bugs of the same class.

In production

For shipping titles with a long support window, watch for this issue resurfacing after dependency updates. Engine upgrades, driver updates, OS releases - each one can resurface a bug class you thought you'd fixed because the underlying behavior changed slightly. Regression tests catch the obvious ones; player reports catch the rest.

When triaging a similar issue in production, prioritize gathering data over hypothesizing causes. A player report describes a symptom; what you need is a build SHA, a session timestamp, and ideally a screen recording or session replay. With those, the bug becomes tractable. Without them, you're guessing at hypothetical reproductions that may not match what the player actually hit.

Performance considerations

If this issue manifests under high load (many actors, many particles, many network connections), profile the post-fix code path with realistic counts. The original cost was a bug; the new cost is real work, and real work has a budget.

Diagnostic approach

Before applying any fix, gather enough context to be confident you're addressing the actual cause and not a similar-looking symptom. The cheapest diagnostic step is reproducing the bug deterministically - if you can't get the same failure twice in a row, your fix attempts will be hard to evaluate. Lock down the reproduction first.

For Godot-specific diagnostics, the editor's profiler is the canonical starting point. Capture a representative frame with the symptom present; compare against a frame without the symptom; the diff often points directly at the cause. If the symptom is non-deterministic, capture multiple frames and look for the pattern - the cause is usually a state transition or a specific input value rather than a continuous effect.

Tooling and ecosystem

Third-party plugins often provide better diagnostics for their own behavior than the engine does. If the affected code is in a plugin, check the plugin's documentation for debug modes, verbose logging, or inspector tools - these can save hours of investigation when they exist.

Within Godot, the relevant diagnostic surfaces include the standard frame debugger, memory profiler, and engine-specific debug overlays. Each one shows a different facet of what's happening. The frame debugger reveals draw call ordering and state transitions; the memory profiler shows allocation patterns; the debug overlay reveals per-system state. Bugs that resist one tool usually surrender to another - the trick is knowing which tool to reach for first.

Edge cases and pitfalls

Edge cases for this class of issue often involve specific timing: the first frame after a state change, the last frame before a transition, frames where multiple subsystems update simultaneously. Reproducing these reliably is part of what makes the bug class hard to test.

When writing a regression test for this fix, focus on the boundary conditions that surfaced the original bug. Tests that exercise the happy path catch obvious regressions; tests that exercise the boundary catch the subtler regressions that look like new bugs but are really the original returning. The latter are the tests that earn their keep over the long life of the project.

Team communication

Document the fix and its rationale in the commit message or attached engineering doc. Future engineers will encounter related issues; the rationale tells them whether your fix is reusable or specific to the case at hand. Without rationale, the fix gets reverted or copied incorrectly.

If this fix touches a system several engineers work in, a short writeup in the team's engineering channel helps. Not a full design doc - a paragraph explaining what was wrong, what's fixed, and what to watch for. Future engineers encountering similar symptoms will search for the fix; making it findable is a small investment that pays back later.

“Stuck agents are a physics problem masquerading as an AI problem. Tune the radius first and the behavior trees second.”

Related Issues

For related character movement bugs, see Fix Godot characterbody move and slide no movement, and for avoidance-style physics issues, Fix Godot characterbody2d jittering sliding walls.

Tip: enable “Show Navigation” in the debug menu and watch the computed velocity arrow — if it points at zero, RVO is why.