cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post

Calendar’s for maintenance windows, SLA’s, and production monitoring windows

Calendar’s for maintenance windows, SLA’s, and production monitoring windows

Calendar’s for maintenance windows, SLA’s, and production monitoring windows

I know this has been beatup before but I am attempting to introduce this at the product core level to be available to all components

in order to increase product capacility, flexibility, and scalability.

Many shops these days varied production hours of operation, 24x7 schedules, maintenance windows and other recurring time periods where monitoring,  alerting, or the absence of applies.  The ability to define a schedule or calendar for various time periods is very helpful.  This calendar can then be applied to a monitor, an alert, a view, or even a report.  By doing so will reduce the number of alerts required to cover windows that vary over the period of a week.

In our environment, for example, we have a series of component monitors and alerts for two servers that monitor for existence of a process or service, MSMQ queue depth, log file update times, etc.  The challenge is that on Monday – Friday from 0200 – 0600 they are in a maintenance window so alerts should not happen.  To further compound this, the window changes on Saturday and again on Sunday to 0200 – 0700 and 0200 – 0800 respectively.  Now our alert count goes way up to compensate.  If the window changes due to business requirements we have to go in and touch a number of resources to adjust allowing for more opportunity to miss something.  To further exacerbate the situation, there are
other servers running related services also affected by this schedule.

So to have one schedule or calendar to define this window but able to be referenced by all of these aspects of the product makes it much easier to set up and maintain.  This results in fewer configuration points to have to touch when making updates therefore reducing chances for errors.  You can set up many different calendars for servers, interfaces, services, processes, monitors, alerts, etc.  This extends the capability of the product a great deal and makes it more scalable and flexible. 

This schedule also needs to be recurring in nature and as a core component available to all of the components of Solarwinds (NPM, SAM, WPM, Alerts, etc.)

Alert Central is not an option for us being it is a pre-rolled VM environment which goes against corporate policy by not being a company standard image.  Therefore it will never be installed here.

The functionality I speak of here is related to what I have seen in Telalert 6e and works very well.  In fact you can call from the command line the tool to see if a calendar is in effect to make use of it externally.

307 Comments

Lots of people at the Chicago SWUG expressed interest in this. There are some folks who are spending lots of time unmanaging nodes before every maintenance window

Wednesday Hump day Bump !

Sweet..glad to see some others at the SWUGs expressing interest.

Keep bumping and sharing folks !

Level 10

Bumping the bump of the hump day bump.

First bump of the  month of June....

"Building out a method on our own," bump...

Jfrazier​ Would this be a valid start?

pastedImage_0.png

pastedImage_1.png

I had actually presented a cheesy method that works at a SWUG here in Dallas a couple years or so ago.

The challenge we have is some of our normal maintenance windows aren't at the same time every day and they

are not server based but rather application based which can span numerous servers. 

My cheesy solution was to set up a scheduled job in the windows scheduler that created a zero length file at a specific time and then deleted at another specific time.

Then I had a file existence monitor for the specific file.

Then in Orion I created steps in the complex alerts to check for the stratus of the file existence monitor as part of the alert logic.  That way the monitors related to an application continued to run and gather data.  When the maintenance window was over the alert would fire if things were out of spec.

This allows for multiple maintenance windows a day for multiple services/servers, etc.

While it would be great if Solarwinds were to build in the logic there is a way to do it within the product.

Pre Hump Day Texas Heat Wave Bump !!!

Last Tuesday of August Bump...

That's some serious syadmin bandaid action