This message was deleted.

# http4k

[Slackbot]

This message was deleted.

[Philipp Mayer]

Here are the DTOs:

private data class RawFlashScoreResponse(
    val DATA: List<League>,
) : List<RawFlashScoreResponse.League> by DATA {
    data class League(
        val NAME: String,
        val EVENTS: List<Event>,
    ) {
        data class Event(
            val FH: String,
            val FK: String,
            val START_TIME: Long,
        )
    }
}

private fun RawFlashScoreResponse.toSchedules() = Schedules(map { Schedule(it.NAME, it.EVENTS.toGames()) })

private fun List<RawFlashScoreResponse.League.Event>.toGames() = map {
    Game(Instant.ofEpochSecond(it.START_TIME), it.FH, <http://it.FK|it.FK>)
}

[dave]

What does the outside layer (invoker) look like?

Ideas:

• If you are getting a 500 then have you tried a simple try catch block?
• better idea - if it's an http lambda put a CatchAll filter around the inbound. And also a debugging filter on the inbound
I'm assuming it's a lens failure - the stack should tell you what went wrong with Moshi

[Philipp Mayer]

I just published my super hacky code here.
Right now, I’m still triggering it via http here.
I set up CatchAll and CatchLensFailure here.

Thank you for your ideas!
I tried a simple runCatching {}.onFailure { printSomething} around the lens, no luck.

The thing I’m suprised about is that it works locally. ( It works on my machine!!!!111 )

[Philipp Mayer]

Here are my logs (debug is set here):

[Philipp Mayer]

Sorry for the spam, but it grinds my gears

[dave]

What do the logs show?
Are you using v1 or V2 of the lambda protocol?
Are you sure that the code doesn't get to the second line? (Just checking!)

[dave]

Reason I ask is that it could be blowing up after its served the request in the aws adapter layer

[dave]

Other thing: same behaviour if you invoke the lambda from the Aws console? Is it running OOM?

[Philipp Mayer]

Here’s the whole log for one triggered request:

[Philipp Mayer]

Using ApiGatewayV1LambdaFunction

[dave]

That's a timeout. Try it again?

[Philipp Mayer]

OOM is checked. Max ~150mb used, 512mb configured

[dave]

Need to see the 500 coming back in the logs 🙃

[dave]

Try upping the timeout or the memory (this increases performance)

[Philipp Mayer]

I’m an idiot

[dave]

We're all idiots occasionally 🤣

[Philipp Mayer]

Double-checking helps a lot, heh. I thought I set the timeout to 70sec. Turns out it was 7..

[Philipp Mayer]

Thanks a lot for your help! Will now continue moving it to the Cron Event. Again, thanks for the great docs and examples everywhere. 👍

[dave]

So does it work then? 🙃

[Philipp Mayer]

It does! The JSON is so big that it obviously takes a little bit longer than 7secs. 😄

[dave]

Yeah, this is why we generally try to go with the generated adapters for Moshi and the Graal runtime. Obvs not for a pet project, but makes a massive difference

[dave]

But - glad you've got it sorted!

[Philipp Mayer]

Well, Graal is step 2! That’s probably the main argument to later on convince my colleagues. 🌚