cURL error 6: getaddrinfo() thread failed to start

I'm running 6 load balanced EC2 servers (Ubuntu) for a PHP 8.1 (Laravel 9) software application. The application uses a ElasticCache-based (redis) queue system to process jobs, and most of these jobs make one or more http calls to various external API endpoints. The application regularly processes hundreds of jobs per minute. I recently upgraded the Laravel framework from 8 to 9. Since then, and especially when queue load is high, my http calls are failing with the error "cURL error 6: getaddrinfo() thread failed to start". Sometimes, this even causes server(s) to go unresponsive for a time.

My main guess is that server or network resources are being exhausted. However, these errors did not start occurring until after upgrading the Laravel Framework, which begs the question of what the application software has to do with resource limits where they seemingly weren't reached previously. No other material functionality was modified in the application during the Laravel upgrade.

Are there any EC2 DNS request limits or fair use policies that could be getting triggered here? Are there Ubuntu limits relating to starting DNS lookup threads? How would I determine if these limits are being reached, and what would be the best way to prevent or mitigate that? Any thoughts, info, and advice is much appreciated!