I've got a scenario once in a while, nearly whenever i blast out a bunch of auto registering active agents i get errors in the gui saying zabbix server is not running with maria/innodb tables. i'll check after the error and everything seems fine and the server log file shows no such crash. I thought this was a bug maybe in 4.2 or maybe 4.4. However using a newer 5.0 with timescaledb and same thing happened. in the gui, but this time in postgres i can actually see a failover happened between my postgres cluster.
I think most likely theres something happening when a mass number of auto registering agents are either overloading one of the various zabbix processes, or maybe there creating/causing too many connections to the database. Does anyone konw what tuning parameters in the zabbix server conf handle thread/processes fo rauto registering agents? also do you know if those processes make db connections themselves and do not go through one of the db processes configured in the zabbix server conf file?
anyone seen something like this?
I think most likely theres something happening when a mass number of auto registering agents are either overloading one of the various zabbix processes, or maybe there creating/causing too many connections to the database. Does anyone konw what tuning parameters in the zabbix server conf handle thread/processes fo rauto registering agents? also do you know if those processes make db connections themselves and do not go through one of the db processes configured in the zabbix server conf file?
anyone seen something like this?