Expose repair messages

[Anbang-Hu]

• TODO 3 in #1182 (email handler - sending email every few hours)
• Repair message for nodes
• Repair message for jobs
• Impacted job count dashboard

TODO

1. Clean up old RepairManager/RepairManagerAgent code
2. @Gerhut Please help expose repair messages for nodes and jobs on dashboard:

For nodes, either /GetVC or /GetVCV2 has two fields like

REPAIR_MESSAGE: "Pending repair by Administrator",
REPAIR_STATE: "OUT_OF_POOL_UNTRACKED",

For a job, /GetJobDetailV2 gives something like

repairMessage: {
  message: [
    "FATAL",
    "The job is running on unhealthy node(s): some_node1(reason1), some_node2 (reason2). Please check if it is still running as expected. Node repair is waiting for the job to finish. Please kill/finish the job as soon as possible to expedite node(s) repair.",
    ""
  ],
  timestamp: 1593127892.838783
}

I'm open to changing message format as you desire.

[coveralls]

Pull Request Test Coverage Report for Build 3548

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 93.438%

---

Totals
Change from base Build 3543:	0.0%
Covered Lines:	818
Relevant Lines:	865

---

💛 - Coveralls

[Gerhut]

Did we got a full repair state list? I got one in https://user-images.githubusercontent.com/5781796/84816053-c7e1b780-afc8-11ea-986f-f09d01b4d78d.png but found OUT_OF_POOL_UNTRACKED not in the graph.

[Anbang-Hu]

Did we got a full repair state list? I got one in https://user-images.githubusercontent.com/5781796/84816053-c7e1b780-afc8-11ea-986f-f09d01b4d78d.png but found OUT_OF_POOL_UNTRACKED not in the graph.

@Gerhut please refer to https://github.com/microsoft/DLWorkspace/pull/1202 for the full graph.