This issue tracks the UI design and mock-up for a new feature which allows you to visualize resource usage within a
single job. This should show how much you are using various resources of the machine. This includes resources such as:
CPU
Memory
GPU
Network
Disk IO
This will be an optional feature and not all service providers will be able to produce numbers for all types of
resources. It is important that the UI can work without it.
We expect that API will be able to deliver updates to a running job and also allow an overview of resource usage after
a job has completed. We expect that the updates will be delivered to the frontend via a mechanism similar to how logs
and status updates are delivered today. For the time being, the frontend should mock these by emitting events to a state
store (e.g. redux or component state) via setInterval.
Similar features from other software include:
Task manager on Windows
Activity monitor on macOS
htop
Graphs as seen on Hetzner and similar cloud providers
For the time being the UI should be hidden behind a hasFeature() flag.
This issue tracks the UI design and mock-up for a new feature which allows you to visualize resource usage within a single job. This should show how much you are using various resources of the machine. This includes resources such as:
This will be an optional feature and not all service providers will be able to produce numbers for all types of resources. It is important that the UI can work without it.
We expect that API will be able to deliver updates to a running job and also allow an overview of resource usage after a job has completed. We expect that the updates will be delivered to the frontend via a mechanism similar to how logs and status updates are delivered today. For the time being, the frontend should mock these by emitting events to a state store (e.g. redux or component state) via
setInterval.Similar features from other software include:
For the time being the UI should be hidden behind a
hasFeature()flag.