Improve messages for events and alerts

[julienlim]

Problem Description

• Messages are too verbose and need to be actionable.
• Alerts should clear when the issue is no longer present.
• Alerts disrupt the user experience and should only be used in important and critical situations, e.g. host down, brick is running out of space, etc.

Recommendations

• Unmanage and Import message should appear as an event (not an alert)
• There should be a Import / Manage cluster event message when cluster is successfully imported. There is none today (in Tendrl 1.6.0)

Current Message:

Cluster 44f8f6f3-a103-437b-a3e1-701db25c5382 moved to un-managed state. The archived monitoring data available at: /usr/share/tendrl/graphite/archive/clusters/44f8f6f3-a103-437b-a3e1-701db25c5382_2018-02-22T19:30:26.102943

Recommendation for Message:

<cluster> unmanaged.  Archived monitoring data to: <node>:<path>

Current Message:

Node node3.localdomain is UP

Recommendation for Message:

<node>  in <cluster_name>  is up

Note: Assumes node is always in a cluster as unmanaged node would not be in a cluster (and not have alert and event messages)

Current Message:

Health status of cluster: gluster-195d43d86fd38ba5929e44529d1fa0b985f42f03946e0bb5ada6999805556674 changed from unhealthy to healthy

Recommendation for Message:

<cluster> is healthy

A more comprehensive list of recommendations can be found at Tendrl Log Messaging Recommendations.

@r0h4n @nthomas-redhat @shtripat @Tendrl/qe @jjkabrown1 @mcarrano

[r0h4n]

@GowthamShanmugam any updates?

[GowthamShanmugam]

@r0h4n i am not started to fix this issue

[GowthamShanmugam]

This issue is fixed, we can close this issue