The test test_validate_tally_on_public_cloud_with_sla is failing intermittently, causing pipeline failures and slowing down the development process.

The test creates a cloud host with a specific SLA and verifies that the tally counts increase by exactly 1. However, it sometimes fails with an assertion error like assert 4 == (0 + 1), showing that the system is counting 4 sockets when we expected only 1.

The test calls reset_account() at the beginning to clean up data, which successfully clears the tally database but doesn't clean up the Host Inventory service. When the test runs sync_tally(), it processes all hosts from the Host Inventory (including leftover hosts from previous tests), not just the newly created one.

Recommended Solutions:
1. Remove the reset_account() call and change assertion from == to >=
2. Fix reset_account() to also clean Host Inventory
3. Use unique identifiers or filters to isolate test data

Acceptance Criteria

• Fix the flaky test test_validate_tally_on_public_cloud_with_sla