management of MQTTqueue overflow!

MQTT queue overflow is not related to script MQTT publishing functions. This error occurs when the CPU is overwhelmed, often due to excessive function calls or for loops, resulting in continuous MQTT overflow messages.

So if you're receiving MQTT overflow messages, it's a strong indicator that your Shelly device is overloaded. To mitigate this, reduce the number of scripts running on your Shelly and the associated script memory usage.

This should help decrease MQTT overflow errors.

Here is a small script, which triggers mqtt overflow messages without using a mqtt publish function.

let stack = [];
function stacking(a) {
stack.push(a);
}
for (let i=0; i<200; i++) {
stacking(['Switch.set', {on: true, id:0}, i]);
print(i);
}

My personal thoughts on this topic:

there isn't an MQTT queue within MQTT itself, it's a specific feature of Shelly devices. My assumption is that there's an internal call queue that drops calls when they time out so not getting processed anymore. This has been happening continuously since firmware version 1.0. In fact, the first beta versions of FW 1.0 used to crash when an MQTT queue error occurred. So, that it now only drops one or two messages is a significant improvement.

Firmware version 1.0 likely requires more hardware resources, especially when all MQTT settings are enabled. The device continuously sends internal MQTT status updates, and if you also use the BT gateway and cloud functions, the device is essentially operating at its hardware limit. This results in numerous internal timeouts. Adding a complex script to this setup pushes the device beyond its limits. Frankly, I'm not sure how they intend to fix this issue.