Watch4Net | EMC M&R | VIPR SRM: When does a metric become inactive?

Products

• VMware Smart Assurance

Issue/Introduction

When does a metric become inactive?

Environment

VMware Smart Assurance - Watch4Net/M&R

Resolution

The criteria for a metric to become inactive is the following;

There is an import_variable_status.task that checks metric values, if the metric value is not updated for a configurable time it will be changed to inactive

Located in Backend/Default/Conf directory 
The import_variable_status.task config is as follows;

<parameters>

<param>

<![CDATA[{"inactive":"14d","vgroups":[{"condition": [{"source":"VMWareCollector"}], "group":["vcenter","devtype","parttype"],"max-inactive":"24h"},{"condition": [{"devtype":"FabricSwitch"}], "group":["device","parttype","datagrp"],"max-inactive":"24h"},{"condition": [{"devtype":"PassiveHost"}], "group":["device","parttype","datagrp"],"max-inactive":"24h"},{"group":["devtype","device","source","parttype","collcfg"],"max-inactive":"24h"},{"group":["devtype","device","source","parttype","datagrp"],"max-inactive":"24h"},{"group":["devtype","device","source","datagrp"],"max-inactive":"24h"},{"group":["devtype","device"],"max-inactive":"24h"}]}]]>

</param>

</parameters>
    

About this task
The installed default settings are as follows:

 

If needed, you can change the schedule and the delays. The following list parses the elements in the configuration file.

<schedule>

Sets the run schedule for the task. The installed setting is every 6 hours.

<schedule
 cron="0 */6 * * *"
 xsi:type="schedule-repeated"
 disabled="false">

 

<param> {"inactive":"14d",

The inactive parameter sets how long a metric can be stale before being changed to inactive when all metrics in a specific peer group are stale (indicating a collection failure). The installed setting makes stale metrics inactive after 14 days of inactivity.

<parameters>
 <param>{"inactive":"14d",...

The assumption is that collection failures are typically corrected within a 14-day period and most installations do not want to see fluctuations in metrics that are caused by collection failures. 
If needed, you can decrease (or increase) this period.

 

"max-inactive": "24h"

The max-inactive attribute defines how long one member of a peer group can be stale before becoming inactive.

When no metric is collected for a single member of a group (and metrics are collected for other members), it is not a collection failure. Instead, it typically indicates a permanent change, such as a deleted or reconfigured component. The assumption is that most installations want to see such a change reflected as soon as possible in reports.

For most installations, the 24-hour delay is optimal. The 24-hour delay accommodates the timing of other processes, such as all collector intervals and database processes.

If needed, you can increase this delay period.

Note:

• Only small installations with correctly timed processes can consider decreasing this period.
  
  
• There are multiple max-inactive attributes. All of them must be changed to the same value to maintain integrity of global reporting.

 

WARNING Do not change the vgroup filter definitions. The vgroups are correctly tooled to work together to accommodate all peer group possibilities for all SolutionPacks.