Openai/6971762b-a4d0-8013-af97-ad0960889ed5

This conversation was exported from ChatGPT

gpt-5

User: What do you recommend to this person on reddit

What do you recommend to this person on reddit

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.