I put a lot of effort in developing a sensor (prototype at this stage) so that I have no blocking measurements (with FreeRTOS threads and queues). So when there is no cloud connection, the measurements get bufferend on the sensor itself (and also timestamped by a battery backed-up RTC).
But when the cloud connection is back, data gets pushed (by calling a function so that values are delivered with the correct timestamp). But now I run into the next problem. I can have a trigger when the functions called but what if in the cloud
- The Webhooks fails
- There is no connection to the storage service
- There is a problem in the Particle cloud
Is there a way (not that I see in any documentation at this point in time) to have that data stored in the Particle cloud until it is confirmed the data got processed by the web service at the other hand of the Webhook? Or will there be in the future?
It is nice to have real-time values, but we need the full chain from sensor to end cloud to have a 100% delivery guarantee (even if it is delivered later).
In our case we want to store it in AWS and have afterwards data analysis and in a second stage real-time forecasts on it so we can send commands back to the sensor or the customer.
But to achieve a 100% fail-safe chain we must have the ability to have (temporary) storage in the particle cloud.
How should this be addressed?
Thank you for any reply on this important topic to us. This is actually for us the last hurdle to take in being able to use Particle in an industrial production ready set-up (first one was to have non-blocking measurements and buffering in the device).