-
제목
How to identify a toplogy churn in the Diagnostic Snapshot file -
설명
The DiagnosticSnapShot file contains a table which lists topology objects and the counts they have hit.
| Topology Type | Num Instance Versions | Max Version | Num Effective | Num Recent Versions | Num Recent Instances |
This article seeks to explain in detail what each column represents, and where to find the Frequent Changes information within the Diagnostic Snapshot.
-
해결 방안
Topology Type – the observed topology object type
Num Instance Versions – total number of changes on any instances of that particular topology type.
Max Version – sum of changes on the most volatile instance. This looks at the instance with the most number of changes and is the total of all changes on that particular instance. This is useful to learn if it is one particular instance or multiple instances.
Num Effective – number of current instances. This value is directly related to the Foglight registry variable 'foglight.limit.instances'. When that registry value's threshold is exceeded, the 'Foglight Topology Size Limit Reached' Warning Alarm fires. The Num Effective shows the total count recorded.
Num Recent Versions – sum of all changes on any instance in the last week
Num Recent Instances - number of instances created in the last week
Num Instance Versions and Num Recent Versions allow Topology churn to be identified. If that number is nearing 5k then there’s a likelihood that the topology is 'churning' which is a contributor of poor FMS performance (seen in FMS Admin Console delays and timeouts)
+---------------------------------+--------------------------------+------------------+-------------------+------------------------------+-------------------------------+| Topology Type | Num Instance Versions | Max Version | Num Effective | Num Recent Versions | Num Recent Instances |+---------------------------------+--------------------------------+------------------+-------------------+------------------------------+-------------------------------+
-
추가 정보
Previous to Foglight 5.5.5 a script that was commonly run was called ‘find-frequent-topo-changes.groovy’
This information is included in the diagnosticsnapshot file of the FMS support bundle.
Search on the keywords "Begin Frequent Changes for Type"
which lists things like: most volatile object, current version, number of recent changes, ect.
In this example, there appear to be two Infrastructure (hostagents) monitoring the same host causing topology churn. The recommendation would be to deactivate and delete one of the two agents to stop the churn. After the churn ceases wait a few minutes for the FMS to return to its normal performance level