Good Morning Friends,
As DBAs, we are often required to set up alert notification emails for various tasks. For example, there might be a requirement to set up notification alerts if a particular backup job fails. A while back, I came across the situation where I was asked to write a script which sent an email alert to a particular group of users if there was an issue with the Transactional Replication. The users were not happy with the alert message generated by the pre-defined alerts introduced by Microsoft in SQL Server Management Studio since they were not able to understand the description of the message – since the body as well as the heading of the email were not easily interpreted. After gathering the requirements, I decided to implement a more functional email alerts system.
Prerequisites:
Approach:
I identified that a table named sysreplicationalerts which is present in the msdb database and contains the entire replication failure alert message. There are around 14 types of alert messages generated for Replication:
STEP 1
Create a table named notification_plants which contains all the subscriber information i.e subscriber ip address, server name, and general name if there is any for that server.
Script for the same is as shown below:
create table notification_plants ( row_id int, plant_server_ip varchar(100), plant_server_name varchar(100), general_name varchar(1000) )
Then populate the table with the respective server ip, server name of the respective subscribers as well as if there is any general name. By general name, consider for example if a subscriber has been created for Newyork location, then in the general name we shall include Newyork.
STEP 2
This step involves creation of a stored procedure named usp_send_replication_alerts which contains the required logic.
First, I created a table named replicationalerts and subscriber_alert_message. I then populated the temporary table named replication_alerts with the data from sysreplication table. This table is present in the msdb database of the SQL Server.
create table #replicationalerts ( replication_alert_id int, [time] datetime, subscriber varchar(100), alert_error_message varchar(1000) ) Insert #replicationalerts select row_number() over(order by alert_id),[time],subscriber,alert_error_text from sysreplicationalerts where datepart(yyyy,[time]) = datepart(yyyy,getdate()) and datepart(mm,[time]) = datepart(mm,getdate()) and datepart(dd,[time]) = datepart(dd,getdate())
When populating the data, bear in mind that only the current date’s data should be populated. i.e If the script is executed on 23 September 2010 then only those alerts encountered on 23 September 2010 should be inserted into the temporary table.
set @i = 1 select @max_count = max(replication_alert_id) from #replicationalerts while(@i<=@max_count) begin select @subscribername=subscriber from #replicationalerts where replication_alert_id = @i select @patindexname=patindex(‘%%’,@subscribername) select @count = @patindexname-1 select @left_subscriber_name = left(@subscribername,@count) from #replicationalerts select @alerterrormessage = alert_error_message from #replicationalerts where replication_alert_id = @i select @alert_message_patindexname = patindex(‘%.%’,@alerterrormessage) select @count_alert_message = @alert_message_patindexname -1 select @length_alert_message = len(alert_error_message) from #replicationalerts where replication_alert_id = @i select @final_count = @length_alert_message-@count_alert_message select @left_alert_error_message = right(@alerterrormessage,@final_count-1) from #replicationalerts select @time = [time] from #replicationalerts where replication_alert_id = @i insert #subscriber_alert_message select @left_subscriber_name,NULL,@time,@left_alert_error_message update #subscriber_alert_message set general_name = a.general_name from notification_plants a inner join #subscriber_alert_message b on a.plant_server_name = b.subscriber_name set @i = @i+1 end
First the value of the variable @i is initialized as 1. Then we determined the max value of the replication_alert_id column present in the temporary table named replicationalerts. While the value of @i is less than or equal to @max_count the WHILE loop continues and fetches the required information as shown above.
Then a temporary table named #alert_message is created which fetches the data from the temporary table named subscriber_alert_message. It fetches only those data which is between 1 hour.
create table #alert_message ( row_id int, subscriber_name varchar(1000), general_name varchar(1000), alert_message varchar(1000) ) insert #alert_message select rank() over(order by subscriber_name,alert_message), subscriber_name, general_name, alert_message from #subscriber_alert_message where datediff(mi,[time],getdate())<=60
The initialized the value of the variable @j as 1 and the WHILE loop continues till the value of the variable @j is less than or equal to the value of @max_count.
set @j =1 select @max_count = max(row_id) from #alert_message while(@j<=@max_count) begin select @mail_subject = general_name+space(1)+’Replication Has Failed’ from #alert_message where row_id = @j select @message = alert_message from #alert_message where row_id = @j select @mail_body = ‘ Hi Team,<br><br> The reason for failure is:<br><br> ‘+convert(varchar(100),@message)+’<br><br> Regards,<br> Administrator ‘
Then send the email to the desired recipients. In order to do this we need to create a Database Mail Profile. More details on How to create a Database Mail Profile can be found in the Books Online.
exec msdb..sp_send_dbmail @profile_name = ‘Profile Name’ ,@recipients =’Email Address’ ,@subject = @mail_subject ,@body = @mail_body ,@body_format = ‘HTML’ ,@importance = ‘High’ set @j = @j+1 end drop table #alert_message drop table #replicationalerts drop table #subscriber_alert_message
That’s it, now you will have a much more user friendly email alerts notification system.
I hope you have enjoyed reading this article. Please do let me know if you have any suggestions regarding a much better approach for the same.
Please do like us on facebook,
http://www.facebook.com/SQLServerGeeks
Thank you,
Satnam
Leave a comment