Feature Request - New alerts and metrics
dsimpson
Posts: 12
Just downloaded the tool as part of the toolbelt and think it is very promising. Here are a few thoughts:
- Alert for all possible SQL Server services with option to determine what is acceptable rather than just checking for service not running. For example if Integration Services should not be running, I want to be alerted it becomes active.
- Buffer cache hit ratio over time.
- % Disk activity by disk volume for time period.
- % Full database or log file - customizable by database
- Alert if database or log file grew with ability to exclude certain databases
- Alert on response time to connect to the database and run a simple query
- Ability to use averages in addition to actual metrics for time-based alerts. For example the tool currently can configure alert if CPU activity is more than 90% for at least 10 seconds. I may also like to configure an alert if average CPU activity is more the 80% for last 60 seconds. This would take into account high fluctuations in CPU usage that may not last more than 10 seconds but still causes the average utilization to go high over a longer time period.
Add support for multiple instances of an alert with custom responses (ie send email or not)
For example on CPU utilization:
High: If CPU activity 90% for at least 10 seconds - send email
High: If CPU activity 90% for at least 60 seconds - send different email
Med: If CPU activity 80% for at least 10 seconds - Just display on client tool
High: If CPU average activity is 80% for last 60 seconds - send email
Med: If CPU avarage activity is 70% for last 60 seconds - just display status on client tool
- In Job Duration have option to specify minimum average run time. If a job normally runs in 5 seconds but takes 10 seconds to run, I do not want an alert.
David
- Alert for all possible SQL Server services with option to determine what is acceptable rather than just checking for service not running. For example if Integration Services should not be running, I want to be alerted it becomes active.
- Buffer cache hit ratio over time.
- % Disk activity by disk volume for time period.
- % Full database or log file - customizable by database
- Alert if database or log file grew with ability to exclude certain databases
- Alert on response time to connect to the database and run a simple query
- Ability to use averages in addition to actual metrics for time-based alerts. For example the tool currently can configure alert if CPU activity is more than 90% for at least 10 seconds. I may also like to configure an alert if average CPU activity is more the 80% for last 60 seconds. This would take into account high fluctuations in CPU usage that may not last more than 10 seconds but still causes the average utilization to go high over a longer time period.
Add support for multiple instances of an alert with custom responses (ie send email or not)
For example on CPU utilization:
High: If CPU activity 90% for at least 10 seconds - send email
High: If CPU activity 90% for at least 60 seconds - send different email
Med: If CPU activity 80% for at least 10 seconds - Just display on client tool
High: If CPU average activity is 80% for last 60 seconds - send email
Med: If CPU avarage activity is 70% for last 60 seconds - just display status on client tool
- In Job Duration have option to specify minimum average run time. If a job normally runs in 5 seconds but takes 10 seconds to run, I do not want an alert.
David
Comments
For your detailed suggestions, which I've logged for consideration in a future release.
Kind regards,
Paul Galwas
Red Gate Software