-
Problem report
-
Resolution: Fixed
-
Critical
-
6.0.19rc1, 6.4.4rc1
-
None
-
RHEl8
Database hardware:
- 64 cores
- 512GB of RAM
- PostgreSQL 15.3
- TimescaleDB 2.10.3
- all-flash array
- Zabbix 6.2
-
Sprint 104 (Sep 2023), Sprint 105 (Oct 2023)
-
1
Steps to reproduce:
Create discovery rule from the master SNMP item having 5 item prototypes, start monitor about 1000 hosts with this template, get 100% busy LLD workers.
item_discovery | 26 GB | 28250169344 item_tag | 11 GB | 11745476608
Result:
3691797:20230622:152622.129 query [txnlev:0] [select ip.item_parameterid,ip.itemid,ip.name,ip.value from item_discovery id join item_parameter ip on id.itemid=ip.itemid where id.parent_itemid in (9541102,9541103,9541104)] 3691797:20230622:152622.130 query [txnlev:0] [select it.itemtagid,it.itemid,it.tag,it.value from item_discovery id join item_tag it on id.itemid=it.itemid where id.parent_itemid in (9541102,9541103,9541104)] 3691797:20230622:152628.856 slow query: 6.726439 sec, "select it.itemtagid,it.itemid,it.tag,it.value from item_discovery id join item_tag it on id.itemid=it.itemid where id.parent_itemid in (9541102,9541103,9541104)" 3691797:20230622:152628.857 End of lld_items_get() 3691797:20230622:152628.857 In lld_items_make()
Expected:
Fast query execution.
