Hi Aviv, Thanks for your reply. We upgrade from 4.4.7 to 4.4.10 and then 4.5.2 We observed it started to increase since we upgraded to 4.5.2. I have to do dwh vacuuming every 2-3 days. Here for example. ---------------------------------------------- omitted output......... [root@manager ~]# dwh-vacuum -f -v SELECT pg_catalog.set_config('search_path', '', false); vacuumdb: vacuuming database "ovirt_engine_history" RESET search_path; SELECT c.relname, ns.nspname FROM pg_catalog.pg_class c JOIN pg_catalog.pg_namespace ns ON c.relnamespace OPERATOR(pg_catalog.=) ns.oid LEFT JOIN pg_catalog.pg_class t ON c.reltoastrelid OPERATOR(pg_catalog.=) t.oid WHERE c.relkind OPERATOR(pg_catalog.=) ANY (array['r', 'm']) ORDER BY c.relpages DESC; SELECT pg_catalog.set_config('search_path', '', false); VACUUM (FULL, VERBOSE) public.host_interface_samples_history; INFO: vacuuming "public.host_interface_samples_history" INFO: "host_interface_samples_history": found 94115 removable, 70244664 nonremovable row versions in 1903718 pages DETAIL: 0 dead row versions cannot be removed yet. CPU: user: 36.72 s, system: 12.91 s, elapsed: 195.78 s. VACUUM (FULL, VERBOSE) public.host_interface_hourly_history; INFO: vacuuming "public.host_interface_hourly_history" INFO: "host_interface_hourly_history": found 126645 removable, 40469226 nonremovable row versions in 482262 pages DETAIL: 0 dead row versions cannot be removed yet. CPU: user: 20.71 s, system: 5.58 s, elapsed: 115.83 s. VACUUM (FULL, VERBOSE) public.vm_disks_usage_samples_history; INFO: vacuuming "public.vm_disks_usage_samples_history" INFO: "vm_disks_usage_samples_history": found 2028 removable, 1672491 nonremovable row versions in 307111 pages DETAIL: 0 dead row versions cannot be removed yet. CPU: user: 4.35 s, system: 3.77 s, elapsed: 51.81 s. ------------------------------------------------------------------------------------------------- We have the plans to switch dwh and grafana to separate vm. Meanwhile we were curious to know the reason of this rapid increase.