Orchestrator resource metrics interface + Docker implementation

[guzman-raphael]

This issue relates to adding compute resource metrics in:

• PodRunAPI trait
• local docker orchestrator implementation
• PodResult (see below)

// model.rs
struct PodResult {
    // ...
    cpu_usage: Vec<(u64, u8)>, // timestamp vs CPU utilization %
    memory_usage: Vec<(u64, u8)>, // timestamp vs memory utilization %
    // ...
}

Would be cool to have a handle to stream the merics

Depends on #18, #12