Service Availability Report seems not accurate for my servic

[source888]

Until now , i think that the causue of availability report not accurate have big chance because of the bug of soft ok not change to hard ok occurs sometimes in the case the host of service is also down ,and this issue still exist in the nagios core version 4.4.5.

In google search ,i found other guy also reported same kind of issue, described in the following link. So all the time duration between the time point soft OK and later the time point of real ok ,these duration are currently calculated as unavailability time, but actually , these duration is already ok time.

https://github.com/NagiosEnterprises/na ... issues/730

[scottwilkerson]

What are you selecting on step 3 when you run the report?

[source888]

Hi, Scottwilkerson

Here's how i generates the report:

First i select a service, then click in "View Availability Report For This Service" ,then got the report page , then i select "last month" ,then click "update" , then i got the report of last month.

Here's a screenshot of the report of one of my service ,this service went into critical on 2020-02-09 12:14:50 ,at 20:00 ,nagios instance have a scheduled restart , and critical status continue , at 2020-02-10 0:0:0 , the status change to OK. From the alert log, i can see on 2020-02-09 12:44:20,this service got a soft ok , but no hard ok until the end of day .

So in real situration , i think this service have been recovered on 2020-02-09 12:44:20 , but availability report think this service was recovered on 2020-02-10 0:0:0 .

[scottwilkerson]

When you access the report this way, some assumptions are made, it sets includesoftstates=no

To modify these you need to run the report from
Reports -> Availability -> Services
Select Service

Then you can choose all the possible settings for the report

[source888]

Hi,Scottwilkerson

With your "formal" method of generating availability report ,now i can get the correct report.Thanks .

One confusion for me , so that soft ok which last 8 or 9 hours long, is it still a bug or an expected status transfer in your opinion?

[scottwilkerson]

One confusion for me , so that soft ok which last 8 or 9 hours long, is it still a bug or an expected status transfer in your opinion?

I believe this is expected.

From the following you will see SOFT RECOVERY is a special case

Service experiences a SOFT recovery. Event handlers execute, but notification are not sent, as this wasn't a "real" problem. State type is set HARD and check # is reset to 1 immediately after this happens.

https://assets.nagios.com/downloads/nag ... types.html

So the actual state type is set to hard from the Nagios memory perspective but it is logged as SOFT to differentiate the fact that notification are not sent in these cases