Critical Cloud-to-Cloud Latency & Alexa Integration Instability

Hi everyone,

I am a telecommunications expert and a long-time user of the BTicino/Legrand Smart system (Living Now / Zigbee-based gateway).

 

I am opening this thread to report a critical and persistent instability with the Amazon Alexa integration that has been ongoing for at least three weeks.

 

The official "Home + Control" skill is currently unreliable. While the devices respond correctly via the native app, the voice control via Alexa is suffering from systemic failures.

 

Technical Analysis of the Issue:
Cloud-to-Cloud Latency: There is a massive increase in RTT (Round Trip Time) between Legrand/Netatmo API endpoints and AWS servers. Commands often exceed the 5-10 second threshold, leading to Alexa’s "Device is not responding" timeout message.

 

Congestion & Peak Hours: I’ve noticed that the system is significantly more fluid during early morning hours. This strongly suggests API rate-limiting issues, server-side congestion, or insufficient resource scaling during peak times.

 

Peering & Routing Issues: Based on my background in TLC, the behavior points towards peering bottlenecks or routing congestion at major exchange points (possibly MIX, NAMEX, or specific AWS interconnects). It seems like the webhooks responsible for real-time status updates are being delayed or dropped.

 

OAuth2 Token Refresh: There are signs of "silent" authorization failures where the sync between the two clouds breaks, requiring a full skill disable/enable cycle to temporarily restore functionality.
I have already tried power-cycling the gateway, performing a full de-provisioning of the skill, and checking local network integrity. The issue is clearly located within the backend infrastructure.

 

The community is becoming increasingly frustrated (as seen in the recent 1-star reviews on the Alexa Skill Store). We expect a smart home ecosystem of this caliber to provide stable APIs and reliable cloud interconnects.
Is the dev team aware of these latency spikes? Are there any planned infrastructure upgrades or DNS/Peering optimizations in the pipeline for December 2025?

 

Looking forward to a technical response.

 

Best regards.