tree fd545add74be8153e0ba813404b2ad322b651cf2
parent 3e0416e6e529afd7475eceae3745da26a6dfb669
author Fabio Ponciroli <ponch78@gmail.com> 1634934093 +0200
committer Fabio Ponciroli <ponch78@gmail.com> 1635188215 +0000

Always update replication status metrics

In the following scenario the replication status metrics become stale:
* Node1 is serving traffic for projectX,replication lag metrics are empty:
  => projectX_Node1_local_version = t1,
     projectX_remote_version = t1,
     node1_lag=0
  => replication lag metric added for projectX-node1
* Node1 is briefly unhelthy and Node2 serves traffic for projectX
  => projectX_Node1_local_version = t1,
     projectX_remote_version = t2,
     node1_lag=t2-t1
  => replication lag metric updated for projectX-node1
* Node1 is back and serves traffic for projectX:
  => projectX_Node1_local_version = t3,
     projectX_remote_version = t3,
     node1_lag=0
  => replication lag metric is NOT updated for projectX-node1,
     since the event is coming from Node1
* Node1 continue serving traffic
  => replication lag metric for projectX-node1 will never be updated
     and it will always be stale

Always update them regardless the origin of the event.
Bug: Issue 15168

Change-Id: Ibe4ff24583ccad7b14640f7b251094c3fd2d8f95