cpu/memory thresholds should be propagated by supervisor

[andreafalzetti]

Is your feature request related to a problem? Please describe

Currently, to display high usage of CPU/Memory we replicated thresholds in gitpod-cli and jb backend plugin. Ideally, supervisor should own the thresholds and propagate them to the clients.

Describe the behaviour you'd like

• hardcoded cpu/memory thresholds are removed from all components
• thresholds are retrieved from supervisor like we retrieve the resource status

Places in the code to change:

• https://github.com/gitpod-io/gitpod/blob/main/components/gitpod-cli/cmd/top.go#L48-L57
• TODO: add link to JetBrains IDE once [JetBrains] Show Workspace CPU/Memory resources in control center #12053 gets merged

Describe alternatives you've considered

N/A

Additional context

[andreafalzetti]

@akosyakov do you think thresholds should be returned inside ResourcesStatusResponse or shall we have a separate endpoint?

[akosyakov]

@andreafalzetti I thought it will be part of

gitpod/components/supervisor-api/status.proto

Line 240 in 698e6bf

message ResourceStatus {

something like

message ResourceStatus {
    int64 used = 1;
    int64 limit = 2;
    ResourceStatusSeverity severity = 3;
}
enum ResourceStatusSeverity {
    normal = 0;
    warning = 1;
    danger = 2;
}

concrete numbers will be implementation details of supervisor