Thank You Wangda Tan and Naga for the comments.
1) I found you have RM node REST response in the zip, did you change any RMNodes related info?
5) Is there any change to cluster metrics web UI?
>> Yes. Available resource were not calculated correctly when labels were used and this was displayed in cluster metrics. I changed that code also CSQueueUtils, hence attached UI snapshot and REST o/p nodes to show the updated Available resources metric.
2) Instead of putting max am to a separated object, I would prefer to put them to existing resourceUsageByPartition instead of introducing a new object. Even though max-am-limit is not usage, but it describes upper bound of usage. Thoughts?
3) And similarly, am-limit should put to queue capacities as well, now max_am_perc is a part of QueueCapacities.
>> I agree that we could use this way, as its also an upper bound usage. So a grouping will come for AM Resource Limit and AM Used Resources (also user based metric). I will see how better I can group these by using existing DAO object classed. Also will look in renaming existing class as mentioned by Naga for better readability in o/p.
4) When label is configured, max-am-percent and max-am-resource should in the above table of the RM UI (above table is for partition-specific properties and below table is for other general propoerties).
>> Yes. This will be fine, i will make necessary changes for same.