ricbra / rabbitmq-cli-consumer

Consume RabbitMQ messages into any cli program
MIT License
247 stars 49 forks source link

added dead letter box + retry count option #19

Open mjolk opened 9 years ago

mjolk commented 9 years ago

Hi,

We are using this project to do some processing. Needing a retry option i created this branch. I will try to clean up the code some more when i have time. Take a look and please comment:)

ricbra commented 9 years ago

Thanks for PR. Looks as a nice addition, nice work. I added some comments.

I'll try to test this branch this week.

ricbra commented 9 years ago

Could you let me know when you applied the changes? I'll see if I can find the time to test it out.

gaspaio commented 8 years ago

Any news on this issue ?

jdulude commented 8 years ago

I'm in need of a DLX with a max retry count. This PR looks like it would do the trick. Any chance of this getting tested and merged? Thanks!

ricbra commented 8 years ago

To be honest I don't think I'm gonna except any feature adding PR anymore on v1 of this lib. It's untested and therefore too dangerous to update.

Good news is I've started on a rewrite which will be tested and should be more feature complete than the current implementation. However its been some time when I've worked on it though its not abandoned. I'm focussing on completing it before the end of this year 😄

jdulude commented 8 years ago

Thanks for the update. Just so I'm clear, are you saying you're rewriting the entire cli-consumer lib, or you're rewriting the retry feature that's in this PR? If it's the entire lib, do you mind me asking why? We will soon be deploying this version to our production servers and it will be serving as a gateway to a fairly critical path of our application. I just want to make sure you're reasons to rewrite aren't performance or stability based.

andrefigueira commented 7 years ago

Hi @ricbra just wondering are still working on V2 or do you think if the conflicts here are fixed it could be merged? In need of this too now, so i don't lose failing messages and don't clog up my queues? Going to be processing billions of records