Monitoring items should use the dtle name as a prefix instead of the host name

[asiroliu]

Description

Monitoring items should use the dtle name as a prefix instead of the host name

Steps to reproduce the issue

1. deolpy dtle cluster

• makesure publish_metrics is true
• set dtle name is some-custome-dtle-name
• set server hostname is dtle-src-1

2. create job
3. get monitoring items 'http://{dtle_server_ip}:8190/metrics'

Describe the results you received

the monitoring item's prefix is server hostname

# HELP dtle_dtle_src_1_delay_time dtle_dtle_src_1_delay_time
# TYPE dtle_dtle_src_1_delay_time gauge
dtle_dtle_src_1_delay_time{task_name="109_117_Src"} 0
dtle_dtle_src_1_delay_time{task_name="117_109_Dest"} 0

Describe the results you expected

the monitoring item's prefix should be dtle name
dtle_some-custome-dtle-name_delay_time{task_name="109_117_Src"}

Output of ./dtle version:**

9.9.9.9-master-1db03e8

Additional information

(e.g. issue happens only occasionally)

Additional details (log, config, job config etc):

[sjjian]

@asiroliu hi~, what's your expected result?

dtle_dtle_1_delay_time{task_name="109_117_Src"} 0?

[asiroliu]

@asiroliu hi~, what's your expected result?

dtle_dtle_1_delay_time{task_name="109_117_Src"} 0?

yes.
I revised the description of the issue. Make my expected value more intuitive.

[sjjian]

方案：

监控项的设计成这样会合理一点：dtle_delay_time{task_name="109_117_Src", dtle_name="some-custome-dtle-name"}

1. 但是由于目前无法获取dtle_name, 因此暂不实现，About Job.ID and Job.Name #722
2. 将 hostname 信息放到 label 中
3. 参考例子：dtle_delay_time{task_name="109_117_Src", host="dtle-src-1"}

原因：

dtle_name 看起来是一个可变的项，应该是一个label，而监控项 delay_time 应该是一个固定的指标，当监控项的key包含可变的值时会导致该监控项无法被唯一确定。例如在配置 grafana 时公式针对每台服务器都需要修改。