Skip Navigation

Improving the development experience worldwide with Event Grid

Posted on July 24, 2018

Program Manager, Event Grid

Since our updates at //Build 2018, the Event Grid team’s primary focus has been delivering updates that will make it easier for you to run your critical workloads on Event Grid. With that in mind and always aiming for a better development experience, today we are announcing the release dead lettering events to Blob Storage, configurable retry policies, availability in all public regions, Azure Container Registry as a publisher, some SDK updates to include all these new features, and portal UX updates! Let’s dig on what all this means to you.

Dead Lettering

Dead lettering is a common pattern in events and messaging architectures that allow you to handle failed events in a specific way. An event delivery may fail because the endpoint receiving the event is continually down, authorization has changed, the event is malformed, or any number of reasons. But obviously this doesn’t mean the event isn’t important and just be thrown away, as every single event carries critical business value. Even when they don’t, it’s useful to track failed events for post-mortems or telemetry.

Dead lettering is now built right into Azure Event Grid, sending those failed events to Blob Storage. Each Event Subscription can have dead lettering enabled, so if delivery of events to its endpoint repeatedly fails over the course of the retry period, the event can be pushed to a Blob Storage container rather than being dropped. And because Azure Blob Storage is able to emit BlobCreated events, you can receive an event anytime something is dead lettered and have your system take action on it.

Take a look at how easy it is to set dead letter location for your events, and start improving the reliability of your solutions today by enabling it on your Event Subscriptions.

Retry Policies

Working in conjunction with dead lettering, you can now specify for how long you want an event to be retried or how quickly you want it to be dead lettered. If you’re building a real-time system that only cares about what happened in the last five minutes and your system goes down, most likely it isn’t either interesting or helpful to have events from a few hours ago delivered when you come back online. Or maybe you know that if your system starts getting overwhelmed by traffic and you reject some events, having them retried is only going to make it worse. You can dead letter those events and pull them from storage when traffic returns to normal.

Retry policies in Azure Event Grid now allow you to specify both a maximum number of retries or a maximum retry time window for each Event Subscription. Once the lesser of those two numbers is exceeded, the event is automatically pushed to the dead letter destination or dropped, depending on what you have configured to happen.
The defaults are 24h (1440min) and 30 times, but you can easily change those settings to your preferred time window and number of retries to make sure you are not missing a single event.

Global Availability

One of the biggest asks we have had over the last few months is to be present wherever your workloads are. I’m incredibly happy to announce that we have been diligently working on deploying to all public regions, and we are now available in the following ones:

  • Australia Central
  • Australia Central 2
  • Australia East
  • Australia Southeast
  • Brazil South
  • Canada Central
  • Canada East
  • Central India
  • Central US
  • East Asia
  • East US
  • East US 2
  • France Central
  • France South
  • Japan East
  • Japan West
  • Korea Central
  • Korea South
  • North Central US
  • North Europe
  • South Central US
  • South India
  • Southeast Asia
  • UK South
  • UK West
  • West Central US
  • West Europe
  • West India
  • West US
  • West US 2

The few remaining regions will become available in the coming months, followed by availability in Government and Sovereign clouds.

Azure Container Registry Events

Azure Container Registry now natively emits events when an image is pushed or when an image is deleted. This allows you to subscribe to these changes and react to them in real time as your container images change, providing a new way to automate your deployment workflows thanks to this integration.

Watch the properties and schema for Container Registry events, and start using them today to save time on your deployment tasks.

SDK and Portal updates

Preview versions of all Event Grid SDKs have been updated to enable all new and preview features in Event Grid, including Retry Policies, Dead Lettering, event delivery schema, input schema mappings, queues as a destination, and hybrid connections integration. Check the latest versions of all SDKs and choose the one that better fits your needs.

The Azure Portal blades for Event Grid have also been updated to make it easier to find your events within the context of a service, as well as more easily manage your event subscriptions. An advanced editor has also been added to the portal to allow you to directly edit the JSON definition of any event subscription directly in the portal for any advanced or new scenarios you have.

We’re incredibly excited about the updates we are releasing today and love all the feedback we have been getting from the community. We want to continue improving updates to Event Grid making it easier to use, and better at supporting your needs and scenarios. Please keep engaging with us by providing your feedback.