I'm wondering if somebody already experimented with a docker container that watches a beanstalkd using phalcon and execute some work and is willing to share some code. Thinking about creating a container that I can easily scale when needed but would love to get some ideas how to setup this image.
Yes I do. For docker I used two containers, once for beanstalkd the other one for the console. This are the images I've used in my docker-compose:
queue:
restart: always
image: phalconphp/beanstalkd:1.10
container_name: beanstalk
ports:
- "11300:11300"
volumes:beanstalk:/var/lib/beanstalkd
beanstalk_console:
restart: always
image: agaveapi/beanstalkd-console
container_name: beanstalk_console
ports:- "5556:80"
depends_on:- queue
env_file:- variables.env
For beanstalk inside phalcon I've written a module, using cli interface for executing the queue workers and integrating the Incubator beanstalk extension.
Sadly, beanstalkd is deprecated project. So I hesitate to use it in a production (new projects) like I used to... still it is very neat MQ broker.
AWS is still proposing beanstalk and it is still working well. What do you suggest for replacing it? I think at the moment, even for new projects, beanstalk remain the best solution for integrating queues in phalcon, but we can start watching and working on wrappers for other systems..
I think we should implement in incubator or in core at least 2 other queues mostly used - redis and rabbit mq.
Rabbit MQ to start with. Any MQ solution which is current - and supported.
AWS is still proposing beanstalk and it is still working well. What do you suggest for replacing it? I think at the moment, even for new projects, beanstalk remain the best solution for integrating queues in phalcon, but we can start watching and working on wrappers for other systems..
@ruudboon How have you structured logging with queues? I think logs are one of the most critical part when using asyncronous queues and they are the weak part of my implementation.
@cosimo I outputting to the stdout that can easily read by docker. At this point I use beanstalk to easily offload heavy work (Generate PDF's). I write the final status (succes/failed etc) in the database. And thnx for the docker code!