87 votes5 comments · Azure Monitor-Log Analytics » Solutions / Packs Gallery and new IP ideas · Flag idea as inappropriate… · Admin →David Schlum commented
Another feature request out there that seems pretty closely related: https://feedback.azure.com/forums/602299-azure-alert-management/suggestions/16792825-there-should-be-a-maintanance-mode-feature-in-app
Suppression of alerts during maintenance now available in Azure Monitor, as part of Action Rules (preview): https://docs.microsoft.com/azure/azure-monitor/platform/alerts-action-rulesDavid Schlum commented
Any update on this feature?
We have a client that uses a SaaS application that has an outage for maintenance at the same time and day every week. We are getting outage alerts during this time even though it's a planned outage.
We'd like to be able to configure "Planned outage windows" where Application Insights either doesn't test the application or it simply doesn't consider it an outage alert. In effect, under the application, I would love to see a "Planned outages" category where I could add one-off or recurring planned outage schedules that could be potentially shared across applications too. There should be an option under that to apply the planned outage to all tests or only selected tests. Additionally, it should have the option to suppress alerts during planned outages or not. In the graphs, I would like to see a differentiation between tests that were showing down during a non-planned outage time and when they were showing down during an outage. I'm thinking something as simply as using orange dots to show web tests that failed during a Planned Outage and red dots for web tests that failed outside a Planned Outage. Similarly, it would be extremely nice to also be able to determine, in the Planned Outage setup, if failed tests during the Planned Outage effect the Availability Percentages displayed. For most people, the SLA percentages are calculated based on "unplanned outages".
I think with the addition of Planned Outages as a first class citizen in Application insights, it could be exposed in many more places so that technicians troubleshooting "problems" in an application could immediately see if whatever they are viewing occurred during a "Planned Outage". In many cases, this would have saved me minutes or hours of troubleshooting when I ended up finally tracking down that something was broken because updates were happening elsewhere in the application dependency chain and not because of something actually being broken.