Platform availability and architecture

Hello Golioth team

We’re currently prototyping and in the works of transitioning onto another board with Zephyr RTOS compatibility. We’re currently running our own REST Api on a Java server with connected MySQL data, as we’ve so far just had one device for testing. We’re looking to design and implement a scalable system for production and rolling out n-number of devices. Each device will roughly transmit data every 2-5 minutes.

I’ve gone through your documentation, but have not been able to find any information on platform up-time or implementation of high availability architecture. For our use-case, it is detrimental that our devices are able to send data at all times with low latency and that each data-point goes through the pipeline, and for any case of down-time due to server outage or overload, the data is handled so no data is lost.

What have you guys implemented to ensure this for your platform? I might have missed the key notes in your docs, so a point in the right direction is much appreciated.

Thanks in advance
Best regards
Hylleberg

Hey @Hylleberg,

Thank you for reaching out!

Pipelines are built on a low-latency stream processing foundation that allows for high-availability delivery to one or more destinations. In the event of downtime of any internal or external systems, messages can be cached and delivered when services become healthy. For enterprise customers, additional guarantees can be provided, such as longer-term caching of all messages.

We’d be happy to share more information and learn more about your use case by jumping on a call. Please feel free to schedule using this link.