-
Incident report
-
Resolution: Unresolved
-
Trivial
-
None
-
2.2.6
There are several slow queries (with many hosts/items/triggers) that are generated from the frontend UI:
SELECT t.* FROM triggers t WHERE (t.triggerid BETWEEN '100100000160762' AND... SELECT t.itemappid,t.itemid,t.applicationid FROM items_applications t WHERE (t.itemid BETWEEN '100100000141384' AND... SELECT t.itemdiscoveryid,t.ts_delete,t.itemid FROM item_discovery t WHERE (t.itemid BETWEEN '100100000646448' AND...
Namely these ultimately manifest as problematic as long-running queries (SHOW PROCESSLIST; Time > 3600s); notably, as they're SELECTs, the front-end timeout is much less (normally 300s or smaller) so the return to the user from the PHP timeouts has already been exceeded.
It seems like there should be some way to prevent these long-running queries (particularly SELECT) if they're no longer providing value or unable to return to the user.
—