New data service settings in bucket REST API

Per PRD

Data Service
All of the following are in the Data Service, set either dynamically (per-Node, per-Bucket, no persistence) with cbepctl set flush_param, or by using an Erlang diag/eval call to add the setting directly to the cluster-wide per-Bucket Config.

For example, to set the percentage of items warmed up before we enable traffic to a Bucket to 15%:

curl -u Administrator:Passw0rd -XPOST http://localhost:8091/diag/eval \

-d 'ns_bucket:update_bucket_props("BUCKETNAME", \

[{extra_config_string, "warmup_min_items_threshold=15"}]).'

Or:

cbepctl localhost:11210 -b BUCKETNAME -u Administrator -p Passw0rd \

set flush_param warmup_min_items_threshold 15

The set of available options is different for the two methods, where cbepctl set flush_param is a subset of the options available via diag/eval (listed in configuration.json).

Neither method is particularly convenient. While cbepctl is documented (therefore supported and considered "safe" as only specific options can be used), it is not persisted, so needs to be re-applied to each node after a restart. The diag/eval method is persisted, and is cluster wide, but should only be used under guidance from Support, due to the high risk of leaving the cluster in an undesirable state.

The proposal is that we consider adding these settings to the Buckets REST API to make configuration easier, safer, and more consistent, for our customers.

Access Scanner
Setting: access_scanner_enabled
Note: Not available in cbepctl

Expiry Pager
Setting: exp_pager_stime

Might be worth discussing sizing/tuning with them.
Warmup
Settings: warmup_min_memory_threshold, warmup_min_items_threshold

Watermarks
Settings: mem_low_wat, mem_high_wat