Make `status` cheaper (statuslite)

The current `fdbcli status` is expensive. Pulling status frequently can create extensive workload to cluster controller, affecting a cluster's performance.

Scenarios: If clients can monitor its cluster's status more frequently, it can better tune its workload based on the cluster's current load. This enables clients to adapt its workload on the cluster's status.

Observations:
1) Each status request does not need to pull all metrics.
2) Important status may be exposed through some fast path.

[This idea was proposed and discussed in FDB Offset 2019.]

Start with the existing fdbcli status command and compare its behavior with the issue's observations about metric collection and fast paths. Define the statuslite scope and validate that frequent status requests reduce cluster-controller workload while still exposing the important status needed by clients.

Written by the indexing model from the issue text.