At New Relic, I worked on Terraform-based onboarding for AWS customers who needed a repeatable, infrastructure-driven way to connect their AWS environments to New Relic for observability. Instead of relying only on manual, UI-led setup, this flow provisioned the required AWS-side integration resources in customer-owned AWS accounts and linked those accounts to New Relic in a more controlled and auditable way. This aligned especially well with stricter enterprise environments where infrastructure changes needed to be codified, reviewed, and reproduced. (docs.newrelic.com)
AWS onboarding for observability can become operationally messy when done repeatedly across customers and environments. Customers needed to connect AWS accounts to New Relic, configure trust boundaries, enable telemetry export, and do all of that in a way that was reviewable and repeatable. Manual onboarding through the product UI worked for some users, but it was less suitable for teams that preferred infrastructure as code, tighter change control, and explicit ownership of everything created inside their AWS accounts. New Relic publicly supports both product-led setup and Terraform-based setup, which reflects this exact split in customer needs. (docs.newrelic.com)
The observability resources were primarily created inside the customer’s AWS account, not in a New Relic-managed AWS environment. That is the core technical point.
For the Terraform-based path, the integration architecture was:
AWS services in customer account → CloudWatch Metric Stream → Kinesis Firehose in customer account → New Relic ingest endpoint → New Relic platform
New Relic’s public docs state that CloudWatch Metric Streams is the real-time metric export mechanism, and that Firehose can be used to forward those metrics into New Relic. (docs.newrelic.com)
The public Terraform/provider examples show the concrete pieces involved:
PUSH and PULL collection modes, corresponding to Metric Streams and polling-based integrations respectively (github.com)This matters because the case study is not about “setting up observability in New Relic’s AWS.” It is about provisioning the customer-side telemetry export and account-linking path so that New Relic can ingest and visualize that data.
I automated the onboarding flow for AWS customers using Terraform so that connecting AWS environments to New Relic required less manual provisioning and less repeated setup work. The work centered on codifying the integration flow, reducing friction in onboarding, and making the process more suitable for customers who preferred infrastructure-as-code-driven operations over UI-led setup. That core claim is directly supported by my resume.
In practical terms, the contribution can be described as: