Github user lamber-ken commented on the issue:

    https://github.com/apache/flink/pull/5857
  
    @zentol , thanks for review.
    first, there is a small point that the IDs may be duplicated when use JM/TM 
actual unique ID to compose the job name of pushgateway, like the picture below.
    
![image](https://user-images.githubusercontent.com/20113411/41587417-c7367d86-73e1-11e8-95fc-ae54dbd63809.png)
    
    second, when config the grafana dashboard, the prefix of the pushgateway 
job name is useful, the uniq ID may not useful, because our metric data 
contains (`jm_id`, `tm_id`).
    
    so, for pushgateway, the jobName is used to distinguish different services.
    for grafana, the prefix of jobName can use to distinguish different 
clusters, we also can just use JM/TM ID and ignore the jobName.
    
    `FLINK-9543` is useful and important. I tried to finished it. 
    Is't ok to use `resourceId` to represent the JM‘ID?
    ```java
    public abstract class ResourceManager ...{
        public static final String RESOURCE_MANAGER_NAME = "resourcemanager";
    
        /** Unique id of the resource manager. */
        private final ResourceID resourceId;
    
    ```


---

Reply via email to