Updates fail but still showing online and no errors

I have been running Blynk 2.0 on a local device for months with no problems. The system is a RPI running the Blynk 2.0 Python community project GIT. The system reads multiple sensors and sends updates via the API - pin updates and color changes primarily.

The problems started in April w/the updates no longer being reflected in the Blynk web portal or the mobile app. The API calls execute in my code without errors or exceptions being thrown. The device still shows as online in the mobile app and on the portal. If I restart the application the updates start to show up in Blynk again. So it seems like this is related to a stale connection that is not throwing socket errors but the listener on the other side of the connection is not processing updates.

I noticed the announcement on throttling of http api calls - were there changes that could have crossed over to how the Python SDK interacts w/the backend services? I’m at a loss why this is no longer working - I have the same code running on a dev machine, with fewer updates or not real data reads of sensors - and it has not shown the problem. So appears to be related to the type of data and volume of updates?

Anyone have ideas how to triage this? It happens intermittently and I don’t have a way to easily reproduce.

How many requests do you send daily per device?

Rough averages of writes to Blynk are per below - these include both VPIN write and set_property calls.

28/min, 1680/hr or 40,320day

Strange. Current limit is 100k requests per day per device. So you should fit.

I found the cause of the problem and wanted to share back with you @Dmitriy. Come to find out the wifi signal to the area where the device is located was weak - was measuring 5Mbps internet access from the RPI running the Blynk app. So I dropped an ethernet line for the device, which is at 100Mbs per switch port, and the problem has not occurred.

What does not make sense is why the Blynk API connection became unstable due to poor network connectivity.

So while I solved my problem, it points at the Blynk transport protocol as having some connectivity/stability issues when there is network layer instability. What can be done to further tune/improve so this does not happen again or for users w/unstable connections in the field - eg. cellular based devices?