Update master to use pulp_ansible 0.5.3

This is addressed in pulp_ansible 0.5.3. In place of the SSO token request a sleep() timeout was added. This is a temporary measure until a more robust solution is developed.

cspealma@redhat.com chousekn- if I am understanding this correctly, I think this ticket should go to the automation hub Jira project–do y'all mind moving to the appropriate project?

A summary of this issue has been sent to akamai-help@redhat.com.

A more specific issue, that may be a place to start, is that Akamai sends the wrong error code. We're receiving a 403 error, which indicates "access to the requested resource is forbidden". The sync machinery interprets this as a security issue, and naturally goes and gets a new SSO token. If Akamai could return a 429, which indicates "the user has sent too many requests in a given amount of time (i.e., rate limiting)", our application would know to back off and wait a period of time before continuing with the sync.

The platform infrastructure owns the Akamai configuration.  The Akamai team will do whatever we ask them once we decide a way forward.

The rate limiting is there for security reasons (denial of service protection) , so we need to balance that vs valid use cases.

We need to find out if its possible to customize the configuration (for example set different threshold for Automation Hub API) , but  I suggest we also look more closely at the Automation Hub sync architecture and exhaust other means of solving this problem before relaxing our WAF configuration, so we are not trading one problem for another.

rhn-insights-aprice rhn-support-lphiri cmoore-insights

 Do we just need to have Calvin reach out to akamai-help@redhat.com ? Or Shawn Iwinski?

Thanks!

-kriedese