Identify head blocker in a blocking chain
It would be quite useful and timesaving if one could identify the head blocker for a blocked process (as done in SSMS's Activity Monitor, Processes tab). Otherwise when get a lot of blocked process alerts, you have to manually figure out the head blocker by analysing all the alerts.
This work was completed in version 5 – apologies for the delay in updating.
If you view a “Blocking process” alert, clicking on the ‘Processes’ tab (next to the details tab) will show the head blocking process, followed by any blocked processes.
If this doesn’t meet your requirements, please let us know how we can improve this alert.
As 5.1 is available, I'm wondering which of the points below applies to the "Planned for 5.1" change?
If this is fixed, it would make this a reason to upgrade.
• SRP-9520: On the alert details pages, the “Performance data” tabs now always cover the full duration of the alert (or, for active alerts, the duration up until the most recent collection time)
• SRP-9596: Decreased CPU and memory usage relating to unbound data retrieval
• SRP-9607: Updated the hardware and performance guidelines documentation to reflect current guidelines
• SRP-10194: Updated the styling on the manual license activation screen
• SRP-10407: Improved the wording in the banner displayed when you add a server, to show a “connecting” message instead of a green “success” message
• SRP-10409: Scheduled page refreshes for the overview pages and Alert Inbox no longer trigger a “Loading…” indicator
• SRP-10422: On the SQL Server instance overview page, under “Server properties”, we now display the friendly version name and version number instead of the full Microsoft version number
• SRP-10452: Improved the labels in the “Blocking process” alert details, to clarify that the duration shown is the total duration for which all descendants have been blocked
• SRP-2856: Fixed a “maxDate < minDate” issue that caused an unhelpful stack trace to be displayed
• SRP-9479: Performance counters are no longer collected as negative for Disk avg. read time. Previously, when this happened, the data was discarded
• SRP-9915: On the Configuration > Monitored servers screen, in the “More actions” menu, the “Remove servers” option is now enabled if you have at least one host machine selected
• SRP-9571: On the alert details page, for the “Job duration unusual” alert, the information displayed in the “User” field now matches the job owner
• SRP-9964: Corrected the breadcrumbs on the alert details page for availability group alerts
• SRP-10028: On the Availability group overview screen, the State and Failover columns are now sorted correctly
• SRP-10090: Fixed an issue where the styling sometimes took time to load properly in Internet Explorer
• SRP-10395: Clicking “Remove overrides” for a customized alert now correctly updates the number in brackets in the “Monitored servers” list
• SRP-10403: Fixed an issue that caused the “Long-running query” alert status to show as “Ended” when the query was still running
• SRP-10402/SRP-10405: Fixed issues that resulted in users being unable to open certain “Long running query” and “Blocking process” alerts
• SRP-10431: Navigating to the login page (when logged out) with an incorrect ReturnUrl no longer causes an error
• SRP-10432: Corrected the error message displayed when you try to add a server name containing an invalid character
• SRP-10447: SQL Monitor no longer throws an exception when you move a server out of a group and into “All Servers”
• SRP-10453: When alert emails are enabled, a successful job run following a “Job failing” alert will no longer generate an alert email with a status of None
• SRP-10457: The installer no longer displays a confusing error message if you try to create a database with the same name as an existing database
• SRP-10461: The database schema update warning in the installer is no longer truncated at 125% DPI
Chris Taylor commented
We see the SQL Monitor queries themselves often blocked by longer running queries. Would be useful if alerts weren't generated in this case either. (e.g. by ignoring a particular account)
Was this ever done?
Oftentimes it happens to be Red Gate SQL Backup that is the Blocked Process. I'd really like to QUASH any Blocked Process Alerts when this is the case... Having to spin my wheels working out that a backup is the problem isn't helping me with REAL issues - if I get a Failed Backup, THEN I'm interested. +3 votes...
Thanks for your feedback. We will investigate this.