ServiceLens troubleshooting - Amazon CloudWatch
I don't see all my logsI don't see all my alarms on the service mapI don't see some Amazon resources on the service mapThere are too many nodes on my service map
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

ServiceLens troubleshooting

The following sections can help if you're having issues with CloudWatch ServiceLens.

I don't see all my logs

How to configure logs to appear in ServiceLens depends on the service.

  • API Gateway logs appear if logging is turned on in API Gateway.

  • Amazon ECS and Amazon EKS logs appear if you are using the latest versions of the X-Ray SDK and the CloudWatch agent. For more information, see Deploying ServiceLens.

  • Lambda logs appear if the request ID is in the log entry. This happens automatically for the situations listed in the following table. For other cases, where the runtime does not automatically include the trace ID, you can manually include the trace ID.

Runtime Method Request ID automatically in log entry?

Java

context.getLogger.log

aws-lambda-java-log4j2

Yes

Java

System.out.println

No

Python

context.log

logging.info/error/log/etc...

Yes

Python

print

No

Node.js

context.log

console.log/info/error/etc...

Yes

dotnet

context.Logger.log

Console.WriteLine()

No

Go

fmt.Printf

log.Print

No

Ruby

puts

No

Not all service map nodes support viewing associated logs. View logs for any of the following node types:

  • Lambda Context

  • Lambda Function

  • API Gateway Stage

  • ECS Cluster

  • ECS Instance

  • ECS Service

  • ECS Task

  • EKS Cluster

  • EKS Namespace

  • EKS Node

  • EKS Pod

  • EKS Service

I don't see all my alarms on the service map

ServiceLens shows only the alert icon for a node if any alarms associated with that node are in the ALARM state.

ServiceLens associates alarms with nodes using the following logic:

  • If the node represents an Amazon service, then all alarms with the namespace associated with that service are associated with the node. For example, a node of type AWS::Kinesis is linked with all alarms that are based on metrics in the CloudWatch namespace AWS/Kinesis.

  • If the node represents an Amazon resource, then the alarms on that specific resource are linked. For example, a node of type AWS::DynamoDB::Table with the name “MyTable” is linked to all alarms that are based on a metric with the namespace AWS/DynamoDB and have the TableName dimension set to MyTable.

  • If the node is of unknown type, which is identified by a dashed border around the name, then no alarms are associated with that node.

I don't see some Amazon resources on the service map

For Amazon resources to be traced on the service map, the Amazon SDK must be captured using the X-Ray SDK. For more information about X-Ray, see What Is Amazon X-Ray.

Not every Amazon resource is represented by a dedicated node. Some Amazon services are represented by a single node for all requests to the service. The following resource types are displayed with a node per resource:

  • AWS::DynamoDB::Table

  • AWS::Lambda::Function

    Lambda functions are represented by two nodes— one for the Lambda Container, and one for the function. This helps to identify cold start problems with Lambda functions. Lambda container nodes are associated with alarms and dashboards in the same way as Lambda function nodes.

  • AWS::ApiGateway::Stage

  • AWS::SQS::Queue

  • AWS::SNS::Topic

There are too many nodes on my service map

Use X-Ray groups to break your map into multiple maps. For more information, see Using Filter Expressions with Groups.