How to avoid alerts when SQL Server AG is failed over (With a solution)
Hello all,
I work with MSSQL AG with 2 nodes. One has been primary since I started working and the other node is secondary. We happened to be in a situation where we failed over the AG and now the secondary has become the primary node. It will stay primary until the next maintenance cycle.
This change brought a new surprise for us. A discovered trigger from the template (MSSQL by ODBC) "MSSQL: AG '{#GROUP_NAME}': All replicas unhealthy" has started sending alerts from the node which was primary previously. I had to add a condition in the trigger expression to confirm if the node is primary ( last(/MSSQL by ODBC/mssql.synchronization_health["{#GROUP_NAME}"])=0 ) and that stopped the trigger from sending alerts from the secondary node (which was primary previously).
Should this check be a part of the low level trigger (MSSQL AG '{#GROUP_NAME}': All replicas unhealthy) and likewise triggers which rely on the node being primary?
I am unsure how much overhead it adds for adding that condition but I like to believe it is super useful for those environments in which failovers between nodes are planned.
I would like to recommend this to be added to the template if it is designed by Zabbix core group.
Thank you,
A-DBA
I work with MSSQL AG with 2 nodes. One has been primary since I started working and the other node is secondary. We happened to be in a situation where we failed over the AG and now the secondary has become the primary node. It will stay primary until the next maintenance cycle.
This change brought a new surprise for us. A discovered trigger from the template (MSSQL by ODBC) "MSSQL: AG '{#GROUP_NAME}': All replicas unhealthy" has started sending alerts from the node which was primary previously. I had to add a condition in the trigger expression to confirm if the node is primary ( last(/MSSQL by ODBC/mssql.synchronization_health["{#GROUP_NAME}"])=0 ) and that stopped the trigger from sending alerts from the secondary node (which was primary previously).
Should this check be a part of the low level trigger (MSSQL AG '{#GROUP_NAME}': All replicas unhealthy) and likewise triggers which rely on the node being primary?
I am unsure how much overhead it adds for adding that condition but I like to believe it is super useful for those environments in which failovers between nodes are planned.
I would like to recommend this to be added to the template if it is designed by Zabbix core group.
Thank you,
A-DBA