Skip to main content

nRF Cloud terminology

TermDescription

Artifact

A file or binary that is uploaded and associated with a release. This can be a symbol file (.elf) or a firmware binary.

Build ID

A unique identifier for a given artifact.

Cohort

A grouping of devices. Typical cohorts are alpha, beta, and prod, but cohorts can represent any grouping, such as a deployment facility.

Coredump

A capture of a device's entire memory space. nRF Cloud generates a trace from the coredump.

Device

An edge device that communicates with nRF Cloud. A device can report events, be updated with new software, and be grouped into cohorts.

Active device

A device that has contacted nRF Cloud within the currently viewed time window, which is usually daily, weekly, or monthly.

Inactive device

A device that previously contacted nRF Cloud but not within the currently viewed time window, which is usually daily, weekly, or monthly.

Pending device

A device that has been created but has not contacted the nRF Cloud service yet. This can happen when devices are created through the API or by CSV upload.

Device attribute

The latest value reported for a device for a custom metric, for example battery_level.

Device versus data searches

Filtering and searching for data is often possible through two different lenses. Data searches target only the data that occurred when specific criteria were met. Device searches return the historical data of a device. For more information, see Device versus data comparisons.

Device ID

The unique identifier of a device, often the device serial. It must be an alphanumeric slug that matches the regular expression ^[-a-zA-Z0-9_]+$.

Device set

A collection of devices defined by a search query.

Firmware

The firmware binary that the device downloads to install. It can be a raw binary or a .zip archive that contains several binaries.

Fleet

All devices in a project.

Hardware version

The hardware revision of the device. Hardware companies typically have several revisions of a product across manufacturing stages (EVT, DVT, PVT), and the hardware version surfaces those differences. You can upload different artifacts in the same release for different hardware versions.

Heartbeat

An hourly report that contains vital information about the health of a device. Most of the data in a heartbeat is user-defined and customizable.

Issue

A logical grouping of traces with the same signature.

Open issue

An issue that has not yet been fixed and remains in the list of issues.

Resolved issue

An issue that has been closed and no longer appears in the list of issues.

MAR (Memfault Archive)

A compressed file format that contains several types of data, for example logs, metrics, and bug reports, that can be uploaded to nRF Cloud for analysis.

Metric

A key-value pair representing a useful piece of information that a device records. Examples include uptime, free memory, battery life, and current power state.

Operational hours

The total number of hours of operation that have been reported to nRF Cloud for a given software version, device, or fleet of devices.

Organization

An entity that contains many projects and users. Typically, an organization maps to a single company.

Project

A logical warehouse for all the nRF Cloud data related to a particular class of device. Contains releases, cohorts, and issues.

Project key

The key included in the header of an HTTP request when sending data to or retrieving data from nRF Cloud servers. The project key is allowed or required only on certain endpoints, and it does not grant access to all API endpoints. For general authentication details, see the nRF Cloud API documentation.

Release

A logical grouping of artifacts that are usually built from the same revision.

Full release

A release that can update device firmware to a specific version from any older version ([* → 2.0]).

Delta release

A release that can update device firmware to a specific version, but only from another specific version ([1.0 → 2.0]).

Release activation

The state in which a release is activated in a particular cohort. Any device in the given cohort can then query and download the release.

Normal release activation

Makes a release available to all devices in the cohort immediately.

Staged-rollout release activation

Makes a release available to devices in the cohort incrementally.

(Metric) report

A collection of metric readings sent from a device for monitoring. Reports can be collected at a fixed interval (see heartbeats). Reports can also contain metric readings specific to a user activity or session.

Each report has a type, which is a string value. The type heartbeat is reserved for the hourly heartbeat reports. Otherwise, the type is user-defined.

For more information, see Metric reports.

Revision

The version-control system revision associated with a release. If you use Git, the revision is the SHA-1 of the commit.

Signature

A unique identifier generated for each trace and used for grouping traces into issues.

Software type

A specific software package that runs on a device in a multi-component architecture. For example, if a toaster has two MCUs, one that handles the main system and another that handles Wi-Fi connections, there are two software types: one likely called main-fw and the other called wifi-fw. A software type has many software versions as updates are pushed to the device.

Software version

A single release for a particular software type. A single software version is associated with one symbol software artifact.

Symbols

The .elf file that nRF Cloud uses to symbolicate crash data.

Trace

A crash or snapshot of the device at a particular moment in time, which

nRF Cloud can later convert to a backtrace. A trace is one of three types:

  • Trace event: captures only the program counter (instruction pointer) and, if available, the link register (return address).
  • Stack trace: captures the memory regions of the stack.
  • Coredump trace: captures most, if not all, relevant memory regions of the system.

Trace event

An event captured when the device crashes or when the SDK is told to capture a trace manually. A trace event contains only the program counter (instruction pointer), the link register (return address) if available, and an optional log or status code.