Summary
Add an end-to-end test validating Logger's InvokeStore-backed log attribute isolation on Lambda Managed Instances, where concurrent invocations are multiplexed into shared execution environments.
This includes the supporting testing infrastructure: a capacity provider construct, shared run-scoped capacity provider stacks, and workflow jobs. Part of #5092 — this issue tracks the Logger portion; metrics and batch are tracked separately by the parent. Implemented by PR #5465.
Why is this needed?
Logger's InvokeStore concurrency code paths are currently exercised only by unit tests with mocked Lambda Managed Instances environments. Nothing proves that log attributes remain isolated between concurrent invocations on the real platform. This work validates that behavior against Lambda Managed Instances, following the prototype findings discussed in #5092.
Which area does this relate to?
Solution
Add a Logger LMI end-to-end suite that forces concurrent invocations to share execution environments and verifies that invocation-scoped log attributes do not leak between requests. Add reusable capacity provider infrastructure, shared run-scoped capacity provider stacks, and workflow jobs needed to provision and run the suite.
Acknowledgment
Future readers
Please react with 👍 and your use case to help us understand customer demand.
Summary
Add an end-to-end test validating Logger's
InvokeStore-backed log attribute isolation on Lambda Managed Instances, where concurrent invocations are multiplexed into shared execution environments.This includes the supporting testing infrastructure: a capacity provider construct, shared run-scoped capacity provider stacks, and workflow jobs. Part of #5092 — this issue tracks the Logger portion; metrics and batch are tracked separately by the parent. Implemented by PR #5465.
Why is this needed?
Logger's
InvokeStoreconcurrency code paths are currently exercised only by unit tests with mocked Lambda Managed Instances environments. Nothing proves that log attributes remain isolated between concurrent invocations on the real platform. This work validates that behavior against Lambda Managed Instances, following the prototype findings discussed in #5092.Which area does this relate to?
Solution
Add a Logger LMI end-to-end suite that forces concurrent invocations to share execution environments and verifies that invocation-scoped log attributes do not leak between requests. Add reusable capacity provider infrastructure, shared run-scoped capacity provider stacks, and workflow jobs needed to provision and run the suite.
Acknowledgment
Future readers
Please react with 👍 and your use case to help us understand customer demand.