Enterprise Dashboard Database high utilization

Products

• Service Virtualization

Issue/Introduction

With DevTest Enterprise Dashboard Database running on Azure DB, it is seen that the DB DTU utilization is staying up at 100% continuously.

Environment

Release : 10.6

Cause

Few shortcomings were found in terms of the generation and processing of events that are sent to ED from Registry. 

Resolution

Raise a support case and ask for resolution provided under internal ticket DE540033.

The solution has 3 parts. Please follow the steps carefully.

1. A patch on Registry Server. This patch is intended to suppress access events of internal used used for VSC Integration. This patch need to be placed in ALL REGISTRY SERVERS

 

2. A NEW ED Patch. This patch is an improvement over the already provided patch with additional log statements and improved suppression logic. This patch is needed ONLY ON ED SERVER

 

3. A NEW VSC Patch. This patch arrests connection leak that led to superfluous number of sessions being created from VSC to those Registries that has a VSE with 0 virtual services

 

Steps to Apply:

 

1. Download the patches from support case.

2. On the VSC Server:

> Stop VSC Service
> Move the existing file "bespin-service-1.7.2.jar" from VSC_HOME\lib folder to a location outside VSC_HOME folder
> Place the patched VSC Jar file with same name at the location: VSC_HOME\lib
> Edit applicaiton.yml file from VSC_HOME folder and make the following change:
logging:
  level: 
    com.ca.ad.sv: DEBUG
         

> Start the VSC Service

 

3. On the ED Server:

> Stop ED Service
> Delete the existing ED Patch "dradis-10.6.0-DE540033.jar" from LISA_HOME\lib\dradis folder
> Place the latest file with same name at LISA_HOME\lib\dradis folder
> Make sure the following properties exist in LISA_HOME\dradis.properties file:
dradis.db.accessEvents.flushWhileSaving=false

dradis.db.accessEvents.discardVSCUserAccessEvents=true

> Make sure the following entries exist in LISA_HOME\logging.properties file:

log4j.logger.com.ca.dradis.service.RegistryService=DEBUG
log4j.logger.com.ca.dradis.BeanProcessor=DEBUG

> Start ED Service

4. On ALL Registry Servers:

> Stop Registry service
> Place the file "patch_DE540033_10.6.0_GA.jar" at LISA_HOME\lib\patches folder. Overwrite the file if already exists with same name
> Add the following entries to LISA_HOME\logging.properties file
log4j.logger.com.itko.lisa.dradis.comm=DEBUG
log4j.logger.com.itko.lisa.coordinator.TestRegistryImpl=DEBUG
log4j.logger.com.itko.lisa.coordinator.UserCache=DEBUG

 

>  As there are total six Registries connected to the ED. Please add below property in five Registry installations( inside local.properties file) each with a different value among 10, 20, 30, 40 and 50. Leave the sixth Registry and it will use default value of 0.

lisa.registry.enterprisedashboard.offsetInterval=

> Start Registry service

 

With the 3 patches in place, please observe the DTU usage on the Azure DB and let us know the outcome.