It looks like, if I pushed a new version of application, there is application downtime and availability is < 100%.
How does the availability computed if I have a planned downtime? If I stopped the application for pushing a new version of application, the availability computation should not be affected.
Alternatively is there any way, I can disable availability monitoring during planned downtime, so that computation of availability monitoring is not affected.
You can achieve this by stopping the test(s) related to the application during the maintenance window. However, you may wish to start using blue/green deployments. https://developer.ibm.com/answers/questions/23610/zero-downtime-blue-green-deployments-using-ibm-devops-services-and-bluemix.html
As long as you don't delete the previous app, your tests will remain.
Availability Monitoring is working on a more natural way to support blue/green deployments (2 apps, 1 set of tests)