We have just upgraded from Zabbix 4 to Zabbix 5.

One of out Items runs a MySQL stored procedure via a UserParameter. This Item is set to execute every hour.

For a couple of the clients the query run time was exceeding the server/agent timeout setting (20 seconds).

On Zabbix 4 this just meant that no result was reported.

However once we upgraded the server to Zabbix 5, we observed that it attempted to re-run the item every (normally) 55 seconds (occasionally 35 seconds but mostly 55).

Since the MySQL instance was still running the original query, as well as each subsequent attempt, this eventually caused the MySQL server to die due to the massive backlog of queries.

It appears that Zabbix Server 5 has a feature (which I cannot find any reference to) that is reattempting problem (in this case timed out) Items, to the detriment of the servers that its supposed to be protecting. 

Can this be stopped, or at least have a config item to stop it, before something like this kills other servers.