What were you trying to do that didn't work?

When many AWS instances are started in the same time, when it can cause something like DDoS to IMDS server, because logic of waiting time has change since RHEL-9. The wait() call is only in rhsmcertd_worker.py and the worker tries to collect metadata from IMDS server and it wait random time, when splay=1 in rhsm.conf.

What is the impact of this issue to you?

When many VMs is started in the same time, then IMDS server will response with 429 and auto-registration will be terminated with error and VMs will not have access to content at all

Please provide the package NVR for which the bug is seen:

subscription-manager-1.30.9-1

How reproducible is this bug?:

Always

Steps to reproduce

• Configure VM to use autoregistration in rhsm.conf (auto_registration = 1, splay = 1)
• Start rhsmcertd.server on any supported public cloud

Expected results

The rhsmcertd wait random time first and then i tries to hit IMDS server

Actual results

The rhsmcertd hit IMDS server first and then it wait random time