AWS News Blog

New CloudWatch Metrics for Amazon Simple Workflow

The Amazon Simple Workflow Service (SWF for short) coordinates tasks and manages state for cloud-based applications. You can focus on your business logic and on your application instead of building complex glue code and implementing your own state machines. Among many other use cases, our customers are using SWF to manage complex video encoding pipelines, migrate applications and business processes from on-premises data centers to the cloud, and to manage applications that combine automated processing and human-powered data validation.

Today we are launching an important addition to SWF. It now collects and publishes a wide variety of workflow and activity metrics to Amazon CloudWatch. You can use these metrics to monitor progress, raise alarms if activities are not completed at the expected rate, or to drive Auto Scaling activities to adapt to an environment where the number of workflows and activities varies over time.

Times and Counts, Activities and Workflows
Some of the new metrics report on time intervals. For example, the DecisionTaskStartToCloseTime metric tells you how long a decision task took to complete after it began executing. Other metrics report on counts. As you might guess, the WorkflowsCompleted metric reports on the number of workflows that completed.

If you are hosting your worker tasks on an auto scaling group of EC2 instances, you can use the StartToCloseTime and ScheduleToCloseTime metrics to initiate scale-up and scale-down operations.

Metrics are published for SWF workflows and activities. The Workflow metrics report on both times and counts, as do the activity metrics.

You can view the metrics in the CloudWatch section of the AWS Management Console. Here are some sample workflow metrics:

And here are some sample activity metrics:

Learn More
You can learn more about these new metrics in the newest version of the Amazon SWF Developer Guide:

Jeff;