v73 Extension.Crash

[MarkusSintonen]

After upgrading to v73 (from v66) our Lambdas start to crash.

Used DD_LOG_LEVEL=trace

DD_EXTENSION | DEBUG | Request to https://api.datadoghq.com/api/beta/sketches took 392ms
DD_EXTENSION | DEBUG | Successfully flushed 0 series and 1 distributions
thread 'main' panicked at library/std/src/time.rs:417:33:
overflow when adding duration to instant
RequestId: xyz Error: signal: aborted
Extension.Crash

Nothing else relevant in logs.

[MarkusSintonen]

Might be also related to DD_SERVERLESS_FLUSH_STRATEGY being set to end but not sure.

[astuyve]

Hi @MarkusSintonen thanks for the note!

Sorry you experienced this. In v67 we introduced a breaking change for the next-generation Lambda extension. It added a bunch of changes as well as a new feature, the ability to race and flush data for long-running transactions, which is useful for getting a live picture of logs, metrics, and spans from a running function. For that we adapted DD_SERVERLESS_FLUSH_STRATEGY: end to take a second parameter which is a time interval. I detailed this more here as we recently discovered this overflow issue – which obviously wasn't intentional.

The quick solution is to simply add a millisecond time value after end, so DD_SERVERLESS_FLUSH_STRATEGY: end,900000 would work for the same behavior.

I've actually already merged the fix in #599 and it'll be out in v74.

I hope this is helpful, and thanks again!
AJ

[MarkusSintonen]

Thanks, adding the second param worked

[astuyve]

Great! I'll close this for now and update you when v74 is released.

Thanks again!

[astuyve]

Fixed in v74