Introducing lightweight high-level device metrics in balenaCloud

Now you can surface and view device performance metrics on the balenaCloud dashboard. This is a lightweight way to get a high-level summary of CPU, RAM, and storage utilization, CPU temperature, undervoltage warnings, and other important metrics on every device within your fleet.

This device metric summary is now available in balenaCloud without any extra work. For our advanced users, access this metadata using our open API.

This feature was brought to you by our inaugural balena Release Party. Our team got together to build this feature within a day. Watch the livestream recording to get into all the nitty gritty details.

---

Why did we build this feature?

We live-built this feature to address some recent customer feedback. Our user experienced device failure due to exceeding storage limits. They didn’t have a real way of seeing this kind of data using our platform, so we decided to fix that.

The solution: Let our users surface critical device metrics and metadata from their balenaCloud dashboard, or give them API access to do it themselves.

Our users can access this metadata to…

• See available device resources across your fleet
• Search for idling or overworked devices across your fleet
• Identify devices that might be at risk of failure due to overheating, drives being full, voltage issues
• Tie specific physical hardware to certain performance data to identify performance patterns
• Have a way to detect if a specific device serial ID is being (re)provisioned

We hope to discover more use cases once our dear customers and users try this feature. 🙂

---

Which device metrics can you see now?

Your balenaCloud dashboard will now show device performance data, all done from the supervisor. Once logged into your account, you can see the data from a fleet-wide perspective:

See device-specific performance data by clicking into a specific device within a new card in the top-right part of the device dashboard:

You can access all of these device metadata endpoints using our API. This lets you port this data to a custom dashboard to monitor the utilization and health of your fleet.

Here’s a more detailed look at all of the metrics you can now see.

• CPU usage: See actual temperature value of the CPU, and get updates when the “bucket” of the cpu usage changes past certain milestones in increments of 20%. (cpu_usage).
• Memory usage: See total memory shown stored as bytes, in two fields (memory_usage/memory_total).
• Storage used/Storage total. View utilized storage and total storage as megabytes in two fields. storage_usage/storage_total
  • We will also be adding the block device name to the field storage_block_device
• CPU temperature: See CPU tempreature in degrees Centigrade in boundaries of 3 degrees, subject to bandwidth saving transfer strategies (cpu_temp).
• Undervoltage warning: Show if a device is undervolting (is_undervolted).
• CPU ID: See the CPU serial number provided by a manufacturer(cpu_id).
• MAC Address: See the specific device’s MAC address (mac_address).

Device metrics in your balenaCloud dashboard are updated every ten seconds.

---

Try it out

Device metadata is available in your balenaCloud dashboard once you update the OS on your device, or provision a new device with the updated OS. If you don’t have a balenaCloud account, set one up— your first ten devices are fully-featured and free of charge.

See your device metrics in the dashboard or try and access it using the balena API. Let us know what you think on our Forums.

Until the next Release Party!