How to use Neural Inverse Tracing in Serverless Functions (AWS Lambda, Vercel, Cloudflare Workers, etc.)
Neural Inverse Tracing is optimized to run in the background in order to avoid adding latency to an application. This can cause some difficulties when tracing in serverless environments like AWS Lambda, Vercel Functions, Google Cloud Functions, etc.
The behavior of the Neural Inverse clients is outlined here.
What is the root cause of this issue?
The root cause of this issue is that serverless functions are typically short-lived and terminate after execution. This means that the Neural Inverse agent, which runs in the background, may not have enough time to complete its work before the function exits. Thereby, events might not be sent to Neural Inverse.
Use flush to ensure events are sent before the function exits
To solve this issue, you can manually flush the events before the function exits. This can be done by calling the flush method of the Neural Inverse client. The flush method is supported across all Neural Inverse SDKs and integrations.
Alternative: waitUntil
Vercel and Cloudflare Workers support the waitUntil method, which allows you to wait for a promise (flush) to resolve before the function exits. Thereby you can return a value from your function and still ensure that events are sent to Neural Inverse.
Alternative: flushAt
Alternatively, you can configure Neural Inverse to flush events individually. This can be done by setting the flushAt option to a specific number of events.
