How to Understand AWS EventBridge Pipe "Failed" & "Timeout" Metrics

Select Language:

If you’re using EventBridge Pipes, it’s important to understand the different metrics that show how your pipe is performing and where it might be running into issues. These metrics help you see what kind of failures happen during the process.

First, there’s the ExecutionTimeout. This shows up when your entire pipe takes longer than five minutes, and it stops running.

Then, there’s ExecutionFailed. This means the whole process failed completely — none of the data got through successfully.

Sometimes, only part of the data batch makes it. For this, you’ll see ExecutionPartiallyFailed. It indicates some items in a batch succeeded, but others did not.

Failures can also happen in specific stages of the pipe. EnrichmentStageFailed tracks failures during the data enrichment step, while TargetStageFailed counts failures once data reaches the target, such as Lambda or Step Functions. If only some target items fail, TargetStagePartiallyFailed helps pinpoint those partial issues.

These metrics are separate and don’t all roll into a single failure count. Instead, they provide detailed insights into where and how failures happen during the pipe’s run. For example, a full failure will increase ExecutionFailed, while issues during enrichment will raise EnrichmentStageFailed, and partial batch failures will be logged with specific partial failure metrics.

Partial failures are common in batch processing, especially with sources like Amazon SQS, Kinesis, or DynamoDB streams. If a batch only partially succeeds, EventBridge Pipes will automatically retry the failed items. For Lambda or Step Functions, you can specify what counts as a partial failure by returning a JSON payload that highlights which events need retries. Typically, this includes a “batchItemFailures” list with the IDs of failed items so that EventBridge knows exactly what to retry.

These performance metrics are sent to CloudWatch every minute, helping you keep an eye on your pipe’s health and troubleshoot problems easily. For more details, you can refer to the official AWS documentation on logging, batching, and monitoring EventBridge Pipes.