Inconsistent event.timestamp behavior when using SENTRY_RELEASE in serverless environments

[lucas-devx]

Environment

• Runtime: AWS Lambda
• SDK: @sentry/node v8.2.0
• Node.js: v18.x
• Region: us-west-2

Problem

When using SENTRY_RELEASE in combination with manual captureException(), the event.timestamp is sometimes skewed by several seconds (occasionally even into the future).

This leads to misordered events in the Sentry UI timeline and impacts alerting thresholds.

Steps to Reproduce

1. Deploy a Lambda with SENTRY_RELEASE set via env.
2. Trigger manual captureException() during cold start.
3. Observe event timestamps.

Expected Result

Accurate event.timestamp aligned with actual execution time.

Actual Result

Timestamp offset (e.g. future-dated by 2–4s).

Notes

Seems related to lazy initialization of release context in async environments. May benefit from await Sentry.flush() or lazy-load guards.

[getsantry]

Assigning to @getsentry/support for routing ⏲️