[NR_EXT] v39 - Timeout on ingestion endpoint causing intermittent log loss + incorrect aws_request_id attribution on retry (v48)

[kingyas123]

We are using the New Relic Lambda Extension (enableFunctionLogs: true, disableAutoSubscription: true) to ship Lambda logs directly to New Relic, with CloudWatch intentionally disabled.

We are experiencing two issues:

Issue 1 — Timeout on NR Ingestion Endpoint

• Intermittent [NR_EXT] timeout errors when the extension attempts to send logs to the New Relic ingestion endpoint
• Results in log loss on affected invocations
• Not happening on every invocation — intermittent

Issue 2 — Incorrect aws_request_id Attribution on Retry

• When the extension retries a failed send, the aws_request_id is incorrectly attributed
• The retried log entry is tagged with the wrong invocation's request ID

Environment

Extension version: v39 (also observed in v48 for Issue 2)
Runtime: .NET
Deployment: Serverless Framework
Config: sendExtensionLogs: true, logLevel: debug
CloudWatch: Disabled via IAM Deny on logs:CreateLogGroup/Stream/PutLogEvents

Expected behavior

• Logs should be delivered to New Relic without timeout failures
• On retry, aws_request_id should be attributed to the correct invocation

Additional context

• Zero log loss is critical as CloudWatch is intentionally disabled
• Upgrading to v48 improves retry logic but does not resolve the aws_request_id attribution issue

Questions

• Is the [NR_EXT] timeout on the ingestion endpoint a known issue in v39? Is it resolved or mitigated in v48?
• Is the incorrect aws_request_id attribution on retry a known bug in v48? Is there a fix or workaround?
• Is there any recommended configuration to reduce timeout frequency or improve reliability when CloudWatch is disabled and only the extension is used?

[Newrelic-Boomi]

https://new-relic.atlassian.net/browse/NR-552072

[aavinash-nr]

Hi @kingyas123 , Could you please try with latest layer from here and let us know if issue still persists.
Thank you.

[kingyas123]

Hi @aavinash-nr , I’ll try with the latest layer(arn:aws:lambda:eu-west-1:451483290750:layer:NewRelicDotnet:51) from the link and will update you if the issue still persists.

[kingyas123]

Hi @aavinash-nr,
After upgrading the Lambda Extension to v51, we have started observing multiple warnings in the logs. Below are the details:

Timeout errors while sending logs

We are seeing timeout errors in the extension when sending logs to New Relic.
The extension retries sending the logs, but we have observed the logs arriving in New Relic so far in the pre-prod environment.
Based on this behavior, we are concerned that in production, where log volume and frequency are higher, some logs may be dropped.

Network errors in the extension

We are seeing repeated network errors while sending data to the New Relic collector:
https://cloud-collector.eu01.nr-data.net/aws/lambda/v1

Max network retries exceeded warning
The following warnings and errors are frequently logged:
[NR_EXT] WARN [agentsend] Max network retries exceeded
[NR_EXT] ERROR Failed to send agent payload for request <request_id>:
error sending request for url (https://cloud-collector.eu01.nr-data.net/aws/lambda/v1)
[NR_EXT] ERROR Periodic cleanup: Failed to send stale agent payload for <request_id>:
error sending request for url (https://cloud-collector.eu01.nr-data.net/aws/lambda/v1)

Could you please help us understand if this is a known issue with v51, and whether these warnings can lead to log loss under high log throughput?

Thank you.