NH-89194: upgrade to upstream 2.8.0

[cleverchuk]

Upgrades to upstream 2.8.0.

[cheempz]

LGTM, would be good to confirm the upstream improvement for lambda instrumentation with our layer!

[cleverchuk]

LGTM, would be good to confirm the upstream improvement for lambda instrumentation with our layer!

@cheempz Any specific item you're thinking of?

[cheempz]

LGTM, would be good to confirm the upstream improvement for lambda instrumentation with our layer!

@cheempz Any specific item you're thinking of?

I'm thinking some checks specifically around a function with the RequestStreamHandler interface, where our layer should reflect the upstream fix... hmm and maybe that a function with RequestHandler interface still works as expected, no regression.

[cheempz]

👍

[cleverchuk]

It looks like there isn't any changes affecting those components in the release?

[cheempz]

It looks like there isn't any changes affecting those components in the release?

i was looking at open-telemetry/opentelemetry-java-instrumentation#11868 which is part of the upstream 2.7.0 release, it seems to have changed the lambda instrumentation.

[cleverchuk]

i was looking at open-telemetry/opentelemetry-java-instrumentation#11868 which is part of the upstream 2.7.0 release, it seems to have changed the lambda instrumentation.

Gotcha. It looks like that only affects users that implement those abstract classes instead of the vanilla lambda ones or uses the wrapper for loading their code.

[cleverchuk]

Looking at the code again, I'm realizing to actually capture different lambda events object attributes like APIGatewayProxyRequestEvent http headers, the user will need to use one of these wrappers for the event they're targeting. It explains why the generic one doesn't do anything other than generating the span.
Not true, they use the similar code path. It looks like context propagation will only work with the wrapper version.