Our alerting system is detecting random resource health issues on production databases hosted on Azure SQL MI

Question

Our alerting system is detecting random resource health issues on production databases hosted on Azure SQL MI

Brett Waymouth Admin 20

Every now and then production sql databases:

At Monday, June 22, 2026 at 12:46:41 AM GMT+10, the Azure monitoring system received the following information regarding your Managed database:

We're sorry, we have identified degradation of your database. Currently, Azure shows the downtime for your database resource with up to 15 minutes delay. The actual downtime might be less than two minutes. We're working to determine the source of the problem.

and sometimes a message that status cannot be obtained.

Pilladi Padma Sai Manisha 10,770 Reputation points Microsoft External Staff Moderator

2026-06-25T02:32:12.9133333+00:00

Hi @Brett Waymouth Admin
We would appreciate it if you could share the required details at your earliest convenience, so we can expedite the troubleshooting process

Answer accepted by question author

Alex Burlachenko 23,330 MVP Volunteer Moderator

hi Brett Waymouth Admin & thx for sharing urs issue here at Q&A portal,

resource Health is seeing short platform-level blips on the SQL MI/database resource.

Those alerts can fire even if the real impact is very short. The message itself says Azure may show downtime w/ up to 15 min delay and the actual downtime may be under 2 mins, so ur alerting can look worse than what users actually felt.

‘Status cannot be obtained’ usually means the health signal/health provider couldn’t read state for that moment. It doesn’t always mean the DB was fully down, but it’s still worth tracking if it repeats. I’d line up the Resource Health timestamps w/ SQL MI metrics: failed connections, connection drops, CPU, storage IO, log IO, deadlocks, workers, and app-side timeout/errors. If the app logs don’t show matching errors, it may just be noisy Resource Health signaling.

https://learn.microsoft.com/en-us/azure/service-health/resource-health-overview

https://learn.microsoft.com/en-us/azure/azure-sql/managed-instance/monitoring-sql-managed-instance-azure-monitor

If this repeats on prod, open an Azure SQL MI support case w/ exact UTC timestamps, database/MI resource IDs, Resource Health event IDs if shown, and app-side error samples. Ask MS to confirm if there were failovers, host events, maintenance, or transient platform issues behind those health flips.

For alerting, I’d avoid paging on a single short Resource Health flip. Better to require sustained DB connectivity errors or multiple health events over a few mins, otherwise u’ll get woken up by Azure sneezing.

rgds,

Alex

&

If my answer was helpful pls mark it and additional thx if u follow me at Q&A portal 
and at my blog

0 comments

1 additional answer

Your answer

Pilladi Padma Sai Manisha 10,770 Reputation points Microsoft External Staff Moderator

2026-06-25T02:32:12.9133333+00:00

Hi @Brett Waymouth Admin
We would appreciate it if you could share the required details at your earliest convenience, so we can expedite the troubleshooting process

Answer 1

Hi @Brett Waymouth Admin
Thankyou for reaching microsoft Q&A!
The alerts you are seeing indicate that Azure Resource Health detected a temporary degradation or was briefly unable to determine the health of the Managed Instance. In many cases, these notifications are generated due to transient platform events, and the actual database unavailability is typically very short (often less than two minutes), even though the notification may be delayed by up to 15 minutes.

To help determine whether these alerts correspond to an actual service issue or a transient platform event, could you please provide:

The Managed Instance name and Azure region.

The approximate UTC timestamp when the alerts were generated.

Whether any application connectivity failures, failed queries, or increased latency were observed during the reported timeframe.

The Resource Health history and any entries under Service Health for the affected period.

If the Managed Instance remained accessible and no application impact was observed, the alerts may have been caused by a brief platform health detection event. However, after reviewing the above details, we can determine whether further backend investigation is required.