Purging Revamp Phase 2

[jrafanie]

Followup to #23394

Work Items:

• Test openshift quotas and consider dropping container quota and container quota item archiving completely, see: Add deleted_on (with index) to container_quotas, container_quota_items manageiq-schema#110 https://bugzilla.redhat.com/show_bug.cgi?id=1504560.
• Verify destroy / delete mechanism for each purging mixin types, part 2:
  • Non-containers

    app/models/audit_event/purging.rb:    include PurgingMixin
    app/models/binary_blob/purging.rb:    include PurgingMixin
    app/models/compliance/purging.rb:    include PurgingMixin
    app/models/drift_state/purging.rb:  include PurgingMixin
    app/models/event_stream/purging.rb:    include PurgingMixin
    app/models/miq_report_result/purging.rb:  include PurgingMixin
    app/models/miq_task/purging.rb:    include PurgingMixin
    app/models/notification/purging.rb:    include PurgingMixin
    app/models/policy_event/purging.rb:    include PurgingMixin
    app/models/vim_performance_state/purging.rb:    include PurgingMixin
    

• update purge_orphans tool to purge orphaned records for above non-containers
• keep_reports has support for two types of values for the same setting: 1) A number of report results per report OR 2) a duration (how many months to keep results for each report). Some reports you may run a lot and some not as often. A number # of report results would help on the former scenario but duration (number of months to keep reports) would work on the latter scenario. Maybe we should split this into two settings?
• Review/change the existing keep_* settings that are possibly wrong
  • keep_hourly_performances and keep_daily_performances being both 6 months seems incorrect. We'll have 4320 hourlies vs 180 dailies per object that's live over the those 6 months. Hourlies for trends reports only seem to go back days. Will need to see how far back chargeback reports using hourlies go back.
• Handle container* purging differently. Their lifecycle is different. We have some tables such as event_streams, vim_performance_states, metrics_* and metric_rollups_* that are purged independently. Only vim_performance_states orphaned rows along with by date. event_streams and metrics* tables only purge by date so a container/container_group/etc. could be have been deleted but we won't purge their events or metrics* for months. Maybe they should be treated like vim_performance_states or maybe we should come up with a new way to handle them.