cancel
Showing results for 
Search instead for 
Did you mean: 

Technical Monitoring - Remote Instance Availability Check

Private_Member_19084
Active Contributor
0 Kudos

Hi experts,

yesterday we had the following issue.

A host was extremly slow. It was an VM-Issue.

But, the host was ping-able, so the host-ping metric didn't fire an alert.

However, it was not possible to make a logon on the system.

But, no one of the "instance logon metrics" did fire an alert.

I did check it. The thing is/was, that all the checks are done by LOCAL DAA.

And as the sytem was so slow, no availability-values had been reported from DAA.

So, we had a system which was not working, but didn't got any alert.

My question is now, I did see that in the template for "Technical Systems", there is a "ABAP-System Remote RFC-Availability" metric.

Which is executed by Solman.

(data collector is "RFC (PULL)" and supplier is "RFC Check from Solman").

I would like to have the same check, on instance level.

But I am not really sure, what I have to enter into the field "DESTINAION".

On system level it is "$T_SYSTEM_DESTINATION$".

Is there a similar value on instance-level?

Thank you

Accepted Solutions (1)

Accepted Solutions (1)

shriramcse
Participant
0 Kudos

Hi Christian,

There is no option to check RFC - Availability for each instance. In Technical System that Metric will Check the RFC Availability using READ RFC Created to Managed system.

In your case you can Create a Metric with CCMS Monitoring

Try with this option

Data Collector Type: RFC (pull)

Data Collector Name:CCMS MTE

For MTE check the below heading Availability of ABAP Instances under the link

Availability and Performance Overview Monitor - Alert Monitor - SAP Library

Regards,

Sriram

bxiv
Active Contributor
0 Kudos

If you had an slowness issue with an instance, why are you trying to leverage availability metrics?  If anything you should be reviewing the performance metrics, for example there is a Dialog Response Time metric with a yellow alert at 2 secs and a red one at 3 secs.

Private_Member_19084
Active Contributor
0 Kudos

Hello Bill,

thats was also what I was thinking, why this metrics did not fire an alert...

But than I did understand, that the system was hanging, so nobody was able to do an dialog-step.

And without any dialog-step, no reponse-time measurements, there fore the metric did jump to grey and did not fire an alert

bxiv
Active Contributor
0 Kudos

There is also the Dialog Standardized Response Time which switches the MTE Class to R3DialogDefLoadTime vs R3DialogResponseTime that the Dialog Response Time metric is setup to use.

Private_Member_19084
Active Contributor
0 Kudos

Ok, I will check this.

But I also would like to have a kid of availability alert, if a logon is notanylonger possible, due to hanging host.

bxiv
Active Contributor
0 Kudos

You can have all those metrics active to collect and alert on all of them if that is what you want.  At the instance level though I'm not aware of a way to test logging in, this assumes that your message server is on the same host.

Also keep in mind that you should also monitor your agents to ensure they stay online as many of the metrics pull values from them.

Answers (1)

Answers (1)

Private_Member_19084
Active Contributor
0 Kudos

Hello together,

I would like to open this thread again.

Today we had a similar situation.


Our DB made a problem, so that no SQL-was working fine.

Logon to the system was working.

So no availability metric was firing an alert.

All performance-metrics did switch to grey, as the are CCMS - and ccms was not available as no SQL was working.

So I didn't get any alert, but the system wasn't working again.

Any ideas, how I can make a kind of a check, which really checks if the system is working or not?
Which is executed from Solman?

Kind regards

bxiv
Active Contributor
0 Kudos

I recently was provided a way to test search/index failure via a job in the system, perhaps a basis cleanup/standard job is what you could try monitoring.

Private_Member_19084
Active Contributor
0 Kudos

Hello Billy,.

but where is the data-collection done?
Because if it is in the managed system, it want work, as you get back 0-values.

And so it jumps to grey.

Kind regards

Private_Member_19084
Active Contributor
0 Kudos

I've opened this thread, which may find a way to execute a FM on local FM.