[Bug]: PHPRedis infinite flushing loop about dd-trace-php HOT 4 CLOSED

Hi @B-Galati,

This is indeed the behavior of Laravel Horizon's supervisors. Each second, they perform a monitor loop which eventually generate redis spans while checking for pending commands.

I agree that this creates a tremendous amount of logs. We should eventually do something about this to change the granularity of the debug logs.

Although it wouldn't fix your issue, what you can do for now is enable DD_TRACE_REMOVE_AUTOINSTRUMENTATION_ORPHANS not to have these spans flushed.

from dd-trace-php.

@PROFeNoM thank you but I don't understand. I am using Symfony, not Laravel and in my case, there are no integrations enabled that relate to these Frameworks.

from dd-trace-php.

All my apologies, I read too quickly and overlooked the fact that you weren't using Laravel Horizon, but symfony messenger.

I'm not familiar with Symfony Messenger, but looking at their source code it looks pretty similar. It indeed seems like the consumer waits 1s before checking for new messages, as you observed.

Assuming you are using a redis receiver, I would then indeed believe that the current behavior is expected.

from dd-trace-php.

We are using RabbitMQ but that's not impossible it's calling Redis for some reasons.

It may be because the container is restarted between each messages but I need to dig further.

Not a bug with Datadog in any case!

Thanks!

from dd-trace-php.