TooManyRequestsError doc questions (doubts)

[guilherme90]

Hello. I apologize for creating this ticket. It is not a bug but a question about using the tool.

I'm reading the doc, I even read the code a lot. However, I have doubts about the example in the documentation. After all, where is this "TooManyRequestsError" class? Was this just an example of how we should work around when a timeout occurs on the consumer?

If yes, then should I consider doing something like

try {
 // something
} catch (e: Error | any) {
    if (error.name === 'KafkaJSNumberOfRetriesExceeded') {
        consumer.pause([{ topic, partitions: [partition] }])
        setTimeout(() => {
            consumer.resume([{
                topic,
                partitions: [partition],
                ...consumerOptions
            }])
        }, 5000)
    }
}

Doc example

consumer.run({
    partitionsConsumedConcurrently: 3, // Default: 1
    eachMessage: async ({ topic, partition, message }) => {
      // This will be called up to 3 times concurrently
        try {
            await sendToDependency(message)
        } catch (e) {
            if (e instanceof TooManyRequestsError) {
                consumer.pause([{ topic, partitions: [partition] }])
                // Other partitions will keep fetching and processing, until if / when
                // they also get throttled
                setTimeout(() => {
                    consumer.resume([{ topic, partitions: [partition] }])
                    // Other partitions that are paused will continue to be paused
                }, e.retryAfter * 1000)
            }

            throw e
        }
    },
})

Thanks

[yohaansunnie]

Thanks for posting the question @guilherme90 , I also have the same question going by the docs.
My work has this use-case and hoping someone could provide the right direction here.

[unematiii]

I believe the TooManyRequestsError is st that your service/consumer should raise when it can't keep up with the workload. Name is just an example. Also, I'd be careful with re-throwing the error, as it counts towards the max retries, eventually causing the consumer to crash.