Solution C is built on wrong assumptions about bref-bootstrap-benchmarks HOT 7 CLOSED

FYI I have asked clarification directly to AWS team members about that, I'll update here when I know more.

from bref-bootstrap-benchmarks.

Are you saying solution C does not save billed time? Or that it doesn't not save the actual run time?

Sorry, been on vacation and got snowed in at the cabin with no internet. Just now getting caught back up.

from bref-bootstrap-benchmarks.

Got confirmation by the AWS team, we are not saving billed time and we are not saving HTTP response time (or Lambda duration time).

So C doesn't seem like an interesting solution at all. It should be marked as obsolete in the README.

from bref-bootstrap-benchmarks.

Aren't we only booting Symfony on a cold start though? Or does the Lambda actually kill our bootstrap and then restart it for each request?

from bref-bootstrap-benchmarks.

Symfony boots on every request.

See the bootstrap, it doesn't do much: https://github.com/mnapoli/bref-bootstrap-benchmarks/blob/master/c/symfony/bootstrap

But then on each request Symfony is booted: https://github.com/mnapoli/bref-bootstrap-benchmarks/blob/master/c/symfony/index.php#L19-L20

What I believed is that by booting Symfony before asking for an event/request we could save that from the HTTP response time. It turns out it's not the case.

from bref-bootstrap-benchmarks.

I was mistakenly looking at Solution D and thinking it was C. Doh!

from bref-bootstrap-benchmarks.

No worries ^^

from bref-bootstrap-benchmarks.