This Hibernate query does have some issues... basically the "0 of 11" means that zero nodes are COMPLETELY unreachable. If they have ANY services that are up, they are not considered down.

We need to enhance the UI so that the meaning is clearer. This issue also causes the confusion in bug .

Ok, check the code and found the "AggregatorStatus".
After all, I think the meaning of dashboard is:

• colors:

  • green: 0 services down

  • yellow: 1 service down

  • red: >1 service down

• numbers:

  • left number: number of nodes that are actually down

  • right number: total number of nodes

I've analyzed the source code from version 1.8.16 and 1.10.0 and found the following:
"NodeDaoHibernate$SimpleSurveillanceStatus.getDownEntityCount()" returns "m_nodeCount - m_upNodeCount". So, m_nodeCount => total number of nodes and m_upNodeCount => number of nodes without any outage, where "getDownEntityCount" returns the number of nodes with problems.

In "NodeDaoHibernate", in SQL statement:
– original source code –
count(distinct case when outages.outageid is null and monSvc.status = 'A' then node.nodeid else null end) as upNodeCount,
– original source code –

Shouldn't be?
– proposed source code –
count(distinct case when outages.outageid is not null and monSvc.status = 'A' then node.nodeid else null end) as downEntityCount,
– proposed source code –

In method "getDownEntityCount", instead of returning "m_nodeCount-m_upNodeCount", we should return m_downEntityCount directly.

Applying the patch, both Surveillance and Dashboard looks very nice but, when I link in the new "1 of 11", there are any node in the list. I'm working in that now.

Patch for class NodeDaoHibernate