Instrumenting a Linux IoT device fleet for service monitoring

By Pilgrim - March 12, 2020

At the recent Embedded World show in Germany, Foundries.io demonstrated integration of a customizable Linux-powered device with DevicePilot service monitoring, showing key parts of a modern IoT stack working together.

Foundries.io provides turn-key services to develop, secure, deploy and maintain an embedded, IoT or Edge product. FoundriesFactory also enables Docker applications to be registered and deployed onto devices automatically.

DevicePilot is a tool for IoT service monitoring, which means turning device telemetry into actionable insights, to improve service delivery. It enables drill-down into issues to find their root cause, business KPIs to be easily built and shared, and workflows to triggered by analysis, all by instantly connecting to data with no-code integrations.

The architecture

Instrumenting a Linux IoT device fleet for service monitoring

The above shows a simplified view of the architecture of a typical IoT device and its cloud counterpart. In the demo a monitoring agent installed in a Docker container on the device sends live telemetry via an MQTT broker to the DevicePilot service monitoring tool.

Telemetry from containers

At the show, the agent collected basic OS vital signs such as % CPU, % mem, and disk and network reads and writes. A finished customer implementation would also send application status such as usage and application-level faults. Any programming language can be used within the container. Container(s) can include any necessary business edge logic to define pre-processing, send-rates etc.  It is also possible to send metadata from the Device Management service in the cloud directly into DevicePIlot, to enable monitoring and management of software upgrades. For example, in order to measure “performance by software version”, or track the rollout of a new upgrade, DevicePilot needs to know which version of software each device is currently running.

Simultaneous integrations via both AWS IoT and Watson IoT clouds were demonstrated (with data flowing out from the Foundries stack to each cloud IoT broker, and then back in from each cloud IoT broker into DevicePilot) - with Google, Azure and Pelion coming soon from Foundries too.

Containers can be auto-started on boot and remotely provisioned. They can also be updated OTA without a system reboot. Or, if the customer prefers they can also build this functionality directly into the OS user space if they do not want to use container technology.  Foundries.io provides the secure firmware and OS software including the Docker runtime. To ensure that communications is secure, a chain of trust is established from firmware to end application, and all updates are signed. Foundries also provides sample containers with typical use cases to get customers started. 

Data into value

Once the raw data is streaming into the cloud, the powerful DevicePilot Service Monitoring features such as filters, calculated properties and KPIs can distil it into meaningful business metrics such as device uptime and service availability. These can then drive and automate business actions, to improve service and reduce operational costs.

The most-likely initial finding is that technical and operational snafus will be revealed, which are affecting uptime and service availability. These can then be quickly identified and addressed to support an efficient, growing connected product business.

DevicePilot can now demonstrate - in several markets - a strong track-record of helping its customers drive-up key business metrics such as device up-time, availability and revenue generation, while driving down operational costs - improving bottom line performance and enabling continued rapid business growth.

Working together

DevicePilot and Foundries.io are independent pieces of the modern IoT stack, working together to make it easier to deploy and run IoT devices at scale - to help you build your successful connected-product business.

Comments

See how DevicePilot can make the difference

 

Industry leaders trust DevicePilot to help them analyse and manage their devices at scale.
Book your personalised demo now and discover how DevicePilot can help you 

  • Get to insight faster 
  • Improve operational efficiencies 
  • Enable business users to find the answers they need 
  • Change behaviour using IoT
Erik in a circle-1

Erik Fairbairn, CEO at POD Point:
Achieved 99% uptime across device estate

"We're totally data driven at POD Point, and if we can answer a question using data then we think that’s the best way - there’s no guesswork and you can use the facts.

Our DevicePilot dashboards have really let us get that actionable insight out of our devices."