Improve messages for events and alerts

julienlim commented 6 years ago

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 commented 6 years ago

@GowthamShanmugam any updates?

GowthamShanmugam commented 6 years ago

@r0h4n i am not started to fix this issue

GowthamShanmugam commented 5 years ago

This issue is fixed, we can close this issue

Tendrl / gluster-integration

Improve messages for events and alerts #566