Documentation for version: v0.2.0
Flotta adds a way to manage all your edge devices from a single point using Kubernetes CRD to handle all kinds of device configurations.
Each edge device can register, pull configuration using the Flotta Edge API and keep syncing the data using a secure MTLS protocol between the device and the Flotta Edge API.
It is also possible to schedule a container workload to run in the device using a Workload CRD. A container workload can be scheduled for an unlimited number of devices.
Device agents takes care of running container workloads, but at the same time, it also takes care of the following actions in the workload to make edge-administration life easier.
Device configuration is a hard part on edge, this is why Flotta trusts RPM-OSTree technology to run updates on the non containerized workloads.
Each device can be configured as a single EdgeDevice CR - where all observability, logs, heartbeats can be set - or can be defined on DeviceSets CR. This grouping CRD allows configuring multiple edgedevices from a single unit, making easier the edge-administration life.
Because all the process is based on Kubernetes CRD, it is compatible with your favourite GitOps tools, so deploy pipelines can be defined across all of your edge devices.
For OSTree based devices, the OS lifecycle is managed by the RPM-OSTree system. One of the major concepts of RPM-OSTree is the transactional upgrade and OS rollback. Flotta uses the RPM-OSTree system to upgrade the OS to the desired image commit-ID. The administrator may define the desired commit-ID either on device-level or on device-set level. After the upgrade is performed, the flotta-agent will verify using greenboot scripts the successful upgrade. In case of a failure, the flotta-agent will rollback the OS to its previous state.