Maintenance of underlying infrastructure/runtime for Azure Event Hub / Service Bus

Question

Hi,

I haven't been able to find anything about this in the Azure Documentation - my understanding is that Azure Event Hub / Azure Service Bus are PaaS services and Microsoft takes care of everything related to the underlying infrastructure, and applicative part (the runtime). Is it with no downtime, or is it covered by Azure SLA (99.95%)?

I trust this is transparent and without downtime as for some PaaS services, this is mentioned in the documentation when there is downtime (ex: Azure MySQL) but I'm looking for official source.

Thanks!

Regards,
François

Answer 1

Hi @François Baronnet ,

Welcome to Microsoft Q&A forum and thanks for reaching out here. I have checked with product team and unfortunately we don't have a similar maintenance document to that of the one you shared.

But product team confirmed that Event hubs is a highly available service and our maintenance should not cause any partition downtime longer than 20 seconds.

SLA for Service Bus:

• For Service Bus Relays, we guarantee that at least 99.9% of the time, properly configured applications will be able to establish a connection to a deployed Relay.
• For Service Bus Queues and Topics, we guarantee that at least 99.9% of the time, properly configured applications will be able to send or receive messages or perform other operations on a deployed Queue or Topic.
• For Service Bus Basic and Standard Notification Hub tiers, we guarantee that at least 99.9% of the time, properly configured applications will be able to send notifications or perform registration management operations with respect to a Notification Hub.

SLA for Event Hubs:
For Event Hubs Basic and Standard tiers, we guarantee that at least 99.9% of the time, properly configured applications will be able to send or receive messages or perform other operations on the Event Hub.

• For Basic and Standard tiers, we guarantee that Event Hubs will be available at least 99.95% of the time.
• For Dedicated tier, we guarantee that Event Hubs will be available at least 99.99% of the time.

Ref doc: SLA summary for Azure services

In case if you notice any transient issues, most of the times, retry layers at clients and gateways handle these transient failures. Maintenance is mostly opaque to clients and transient failures should not break the SLA.

Hope this info helps.

----------

Please don’t forget to Accept Answer and Up-Vote wherever the information provided helps you, this can be beneficial to other community members.