LAVA support remote labs with shared boards
We currently have a remote lab (QCOM) that has boards attached and we would like to share the board resources between developers and LAVA because are MTP's and few available.
We already implement a set of commands to get console, power on/off the device, when LAVA request connect_device will lock the board until the LAVA jobs ends, allowing to power on/off the board when needed. With current implementation if a user are using the board the connect_device timeout will trigger and LAVA mark it as InfraestructureError that's fine for jobs but not for health-checks because will set the board in bad status.
An idea to handle this is to allow health-checks retries specified by the user and time between retries. If the board is in bad status will be responsibility of the remote worker admin, in this case us.