DX Unified Infrastructure Management

Hi,

I`m trying to create an availability report in % of our robots / servers to be monitorized.

As I`m looking for options, I thought of installing the net_connect probe on the central hub, and create profiles for pinging every robot installed.

But I realized that, in order to create a SLA/SLO for having availability for every server, I will have to include them manually (hundreds of them).

So I will keep investigating if there`s some automation for this, and will also take a look at unified reporters, but if you have some tip on how to create an availability report with some bulk process (including several servers at once) I will appreciate it.

As an example, we are trying to generate a report of this style:

-SERVICE-- --STATUS-- --AVAILABILITY-- --DOWNTIME--

Server1 OK 99.89% 0:45

Server2 OK 99.98% 0:07

Server3 OK 100.00% 0:00

Did you check if the SLA wizard in SLM suits your needs? It's pretty limited, but it can help you create multiple SLAs at once.

Other than that, there's no bulk configuration / automation available out of the box. I've built a probe to automate SLA creation based on config files. It creates the SLAs directly in the database. I've built it with "general use" in mind, but currently I've developed it more to suit our other integrations. I'm intending to release it one day in the marketplace, but I suspect that's still months away.

-jon

Thanks for the advise Jon

I used the wizard and now I have the SLA based on the net_connect probe as originally intended.

Now figuring out how to export data in a similar way like this:

-SERVICE-- --STATUS-- --AVAILABILITY-- --DOWNTIME--

Server1 OK 99.89% 0:45

Server2 OK 99.98% 0:07

Server3 OK 100.00% 0:00

Actually the customer request us to have direct access to some sort of graphics like this, so still investigating if its possible.

The SLAReports portlet is your friend there. It'll display "downtime" also on the level of periods: when it was down and for how long. The only shortcoming for it is, that it works an SLA level, you can't display information for multiple SLAs at once.

To create a SLA table you can query the database directly. I've posted some examples in my blog, I believe they should be helpful if you need to go down that road: Some helpful SQL scripts for SLA management

SLAreports portlet!!! that`s it!

And I can even export the reports to PDF.

This is exactly what I was looking for.

Now with the SLA wizard + the SLAReports portlet, I can achieve what I was looking for.

I will also take a look to your queries, which are really useful.

Thank you! you rock!

Hello Nicol,

We are using netconnect probe here. So incase there is a disconnection between Hub and robot , the robot buffers the data. Niow in the report for this period does it reflect the availibility as downtime or as the robot buffers the data it is not counted as downtime?

What about using uptime metric using cdm for this report? which is preferred?

Hi Kalyan,

I believe when you use CDM as an option, everytime when the server reboots it is going to reset the data to zero. That is one of the reasons where I didn’t use cdm. I would still prefer using the net_connect by that way you can also avoid dependencies due to any issues on the robot front. With net_connect however, is pretty straight forward.

-kag

Agree Kanam.

But with using netconnect, let's say there is a network issue between the net connect central hub and the robots with robots still staying up. Does this include as a downtime in sla reports?

By the way i am new to sla reports. Do you have customized guide on how to configure this using net connect?

Hi Kalyan,

In that case, you would have received a ping down alert and also, in this case I believe this will be a data loss.

And  there isn’t any dedicated documentation as such to configure SLA reports using net_connect. But you can give the polling interval in the place of “Less than or equal to” under ‘Expect Quality of Service to be’ block.

-kag

Please refer the below.

SLA Reports - CA Unified Infrastructure Management - 8.1 - CA Technologies Documentation

But by doing this way, wont this create multiple SLOs one for each host? Rather being in one single SLO hosting multiple hosts?

-kag

Correct Kag, but currently it fits our needs and it seem to do the job (Specially when you are working with one origin --> net_connect installed on one robot, pinging many hosts).

I also create an SLA for every server with only one SLO in it. It's really all rather philosophic.

Thanks Jon. However, when you are going to publish this as a report won’t we have some challenges? We might need to pull multiple reports? Or is there a way that we can make use of it?

-kag

Yes, thats the issue with slareports