When NEW_RELIC_APM_LAMBDA_MODE is true, the NEW_RELIC_APP_NAME environment variable controls how your Lambda functions appear as APM entities in New Relic. This page explains the key behaviors to understand before setting this variable.
Entity behaviour
Functions sharing the same NEW_RELIC_APP_NAME value and New Relic account are grouped under a single APM entity in the New Relic UI.
If the name you set matches an existing APM entity — for example, one already created by a non-Lambda service — telemetry from your Lambda function merges into that entity automatically. There is no separate entity created.
重要
Changing NEW_RELIC_APP_NAME creates a new APM entity. Historical data associated with the previous name is not migrated to the new entity.
Multi-region reporting
Multiple Lambda functions deployed across different AWS regions can share the same NEW_RELIC_APP_NAME. All of these functions report into a single APM entity, regardless of which region they are deployed in.
To distinguish individual functions or region-specific deployments within the grouped entity, filter by the tags.aws.arn attribute in NRQL or use the tag filter bar in the entity view.
Cross-runtime reporting
Lambda functions using different runtimes (for example, a Python function and a Node.js function) that share the same NEW_RELIC_APP_NAME create separate APM entities — one per runtime. The grouping does not cross runtime boundaries.
Same app name across Lambda and non-Lambda services
If a Lambda function and a non-Lambda APM service (for example, a traditional web application or microservice) are both configured with the same NEW_RELIC_APP_NAME, they report into the same APM entity.
This can be intentional: if you want a unified view of a service that runs on both Lambda and server-based infrastructure, sharing the name allows all telemetry — transactions, errors, throughput — to appear in one place.
However, if done unintentionally, it can cause confusion. Metrics, error rates, and throughput from the Lambda function and the non-Lambda service are combined, which may make it harder to diagnose issues or understand performance independently.