HTTP interface monitoring optimization

[xxo1shine]

There is a memory leak problem in the getstatsinfo interface.

net.api.detail can add unlimited records, when the node enables the metrics function, it is easy to be attacked.

Problem recurrence:

Execute the following command.
curl -X POST http://127.0.0.1:8090/xxx

Then execute the following command to get the metrics information.
curl -X GET http://127.0.0.1:8090/monitor/getstatsinfo

It can be found that there is one more record in net.api.detail.

                {
					"failQps":{
						"count":1,
						"fifteenMinuteRate":0.0,
						"fiveMinuteRate":0.0,
						"meanRate":0.3301391581846841,
						"oneMinuteRate":0.0
					},
					"name":"/xxx",
					"outTraffic":{
						"count":0,
						"fifteenMinuteRate":0.0,
						"fiveMinuteRate":0.0,
						"meanRate":0.0,
						"oneMinuteRate":0.0
					},
					"qps":{
						"count":1,
						"fifteenMinuteRate":0.0,
						"fiveMinuteRate":0.0,
						"meanRate":0.33013738991015756,
						"oneMinuteRate":0.0
					}
				}