Forum Discussion

Tisha_Levy's avatar
Tisha_Levy
12 years ago

Thresholds - Customized thresholds per datapoint based on host group or dynamic ''Applies to''

Hello Andrew,rnAs always appreciate the input! Your scenario is not an uncommon need and that is why we already have that ability built in. You can set different thresholds based on host groups, specific hosts and even down to the instance level. This help article should prove helpful in detailing the appropriate steps to set this up.: http://help.logicmonitor.com/using/i-got-an-alert-now-what/how-do-i-adjust-thresholds/ or contact us directly via Chat with Engineer button or emailing support@logicmonitor.com for any further assistance.

rnIn addition, if you have your different host groups broke out with those specific needs, you may want to consider dynamic groups when adding more hosts (to inherit properties, etc). This article outlines the options and benefits of this as well: http://help.logicmonitor.com/using/managing-hosts/managing-groups/.rnHope this helps in the interim and please let me know should you require any further information or assistance on this matter.rnHave a great dayrnTisha

2 Replies

Sort By
Replies have been turned off for this discussion
  • Andrew_McAllist's avatar
    Andrew_McAllist
    12 years ago

    Challenge:

    One of the challenges with our application is that certain layers of our application need different thresholds. CPU at 50% may be fine for one Component but bad for another.

    To address this, we have mirrored and copied many of the Main Datasources like NetSNMPCPU and added \'\'applies to\'\' parameters to create unique CPU datasources per layer of our application.

    This works but introduces additional administrative overhead when tweaking datapoints or thresholds universally for all NetSNMPCPU datasources.

    Feature request:

    Allow alert threshold wizard to have a column to designate different thresholds for different filters or different groups of hosts. This will allow customized alerting and thresholds while having only one NetSNMPCPU datasource.

  • Andrew_McAllist's avatar
    Andrew_McAllist
    12 years ago

    Ahh Ic,rnThanks for pointing that out. Another reason for having individual data sources for same metric is in certain situations, adjusting datapoints or snmp Poll times will blow away all historical data for that datasource. Some times we need to make adjustments such as this but only want to impact the Layer of the application that requires the change so its not destructive to the data for a datasource that applies to All or multiple layers.rnBut the above does make things easier,rnThanks,rnAndrew