Editing Openai/6971762b-a4d0-8013-af97-ad0960889ed5 (section)

=== Assistant: Here’s what I’d recommend to them—framed in a practical, Reddit-appropriate way: ===
Here’s what I’d recommend to them—framed in a practical, Reddit-appropriate way:

That Status.BUFFER_FULL (17) error is almost always Zigbee network congestion or coordinator overload, not Home Assistant itself.

Based on what you described (≈65 Zigbee devices, ZHA, PoE SMLight coordinator), I’d suggest this checklist:

===== 1. Don’t reboot HA to “fix” it =====

Rebooting clears the queue temporarily, but it hides the real issue. If this keeps happening, the Zigbee stack is getting saturated.

===== 2. Check Zigbee router vs battery balance (this is the big one) =====

If too many of those 65 devices are battery-powered, you’ll hit problems.

Rule of thumb:
* 1 powered Zigbee router for every 6–10 end devices
* Routers = smart plugs, in-wall switches, bulbs (powered)

If you’re light on routers, add a few Zigbee plugs in problem areas and let the mesh settle overnight.

===== 3. Audit automations for message storms =====

Recently added scripts + motion automations can flood the network.

Common offenders:
* Motion sensors firing → multiple light calls
* Repeated turn_on commands even when lights are already on
* Loops or automations without debounce

Fix:
Add conditions like:
* if light is off
* mode: restart or mode: queued (with limits)
* Short delays (even 200–500 ms helps)

===== 4. Verify Zigbee channel vs Wi-Fi overlap =====

If you’re on:
* Zigbee 11 + Wi-Fi channel 1 → ❌ bad
* Zigbee 15/20/25 → usually safer

Once chosen, don’t change channels unless necessary, and let the mesh heal for 24 hours.

===== 5. Update coordinator firmware (but only once) =====

The SMLight PoE coordinators are solid, but make sure:
* Firmware is current
* No experimental builds

After updating, leave it alone—constant resets hurt mesh stability.

===== 6. Watch ZHA logs before it breaks =====

When this happens again, look for:
* Message retries
* Delivery failures
* Devices spamming state updates

That usually points straight to the problematic device or automation.

===== 7. Capacity-wise: you’re not “too big” =====

65 devices is fine for ZHA if:
* Mesh is healthy
* Enough routers
* Automations are sane

Most “Zigbee limits” people hit are actually design issues, not hard limits.

TL;DR:
Add more Zigbee routers, reduce automation spam, don’t rely on HA reboots, and let the mesh stabilize. BUFFER_FULL means “too much too fast,” not “you’re out of capacity.”

If you want, I can also help them:
* Identify router gaps
* Sanity-check an automation
* Interpret ZHA logs

That would be my Reddit reply.