Announcing the Saviynt Knowledge Exchange unifying the Saviynt forums, documentation, training,
and more in a single search tool across platforms. Read the announcement here.

Analytics v2 - gateway time-out

flegare
Regular Contributor III
Regular Contributor III

We built an analytic for SIEM integration and while it worked great for a few days it has been sputtering lately and we now receive gateway time-out more ofen than not.

Here is the SQL statement used:

select ua.TYPEOFACCESS as 'Object Type',ua.ActionType as 'Action Taken',u.username as 'Accessed By', ua.IPADDRESS as 'IP Address',ua.ACCESSTIME as 'Event Time',ua.DETAIL as 'Message' from users u , userlogin_access ua, userlogins l where l.loginkey = ua.LOGINKEY and l.USERKEY = u.userkey and ua.AccessTime >= (NOW() - INTERVAL ${timeFrame} Minute) and ua.Detail is not NULL

I don't see anything especially nasty in there.  The same analytic in v1 is responding without any issues.

I deleted the v1 analytics thinking there might be a conflict but it does not seem to help

6 REPLIES 6

flegare
Regular Contributor III
Regular Contributor III

Request body:

{
    "analyticsname""SIEM Integration v2",
    "max""1",
    "offset""0",
    "attributes": {
        "timeFrame""1"
    }
}
 
Tried with id instead of name, with the same behavior
{
    "analyticsid""1499",
    "max""1",
    "offset""0",
    "attributes": {
        "timeFrame""1"
    }
}

flegare
Regular Contributor III
Regular Contributor III

I did find this in the logs, which could explain the issue:

 

2023-05-05T17:26:35-04:00-ecm-2023-05-05T21:26:35.351667674Z stdout F 2023-05-05 21:26:35,351 [http-nio-8080-exec-269] DEBUG services.SaviyntGuardRailService - Heap memory threshold exceeded.
 
2023-05-05T17:26:35-04:00-ecm-2023-05-05T21:26:35.351698774Z stdout F 2023-05-05 21:26:35,351 [http-nio-8080-exec-269] DEBUG services.SaviyntGuardRailService - Retry, waiting for 60 seconds before resuming the execution...
 
I'll get the instance restarted, it could help...

what is count of analytics records


Regards,
Rushikesh Vartak
If you find the response useful, kindly consider selecting Accept As Solution and clicking on the kudos button.

flegare
Regular Contributor III
Regular Contributor III

It depends on the number of available items in userlogin_access ua and userlogins for the given timeframe.

flegare
Regular Contributor III
Regular Contributor III

Restarting the service appears to have greatly helped, I will keep monitoring

flegare
Regular Contributor III
Regular Contributor III

Marking this as answered.  Service restart has proved to be the solution