Add an option to make a persistent interceptor queue #57
Conversation
coyote.go
Outdated
| ) | ||
| if err != nil { | ||
| return fmt.Errorf("%s %w", color.RedString("failed to declare a queue:"), err) | ||
| } | ||
| if persistent { | ||
| defer func() { | ||
| log.Printf("💔 Deleting queue %s", q.Name) |
There was a problem hiding this comment.
This will delete the queue even if it is persistent right? Can you help me understand what is the benefit here?
There was a problem hiding this comment.
On a second look, I have an idea to use --queue argument to create persistent queue. Allow coyote to reconnect to an existing queue.
By default we generate interceptor queue names randomly with an UUID suffix:
<queue_name:interceptor>.<uuid>
We can make it persistent if queue argument is provided.
# queue name will not be suffixed with an UUID.
coyote --queue <queue_name>
In this case coyote will try to bind to the queue if it exists or create it does not.
So persistent flag is not needed anymore. What do you think?
There was a problem hiding this comment.
You right! The logic here is coupled with a reconnection logic - if the queue is not persistent and we're disconnected from a cluster then RabbitMQ will automatically delete it.
Do you think it might be useful to have an extra flag to enable/disable the queue auto deletion?
There was a problem hiding this comment.
For now I tied it to queue argument being present or not. Let's revisit with auto reconnect functionality.
There was a problem hiding this comment.
That's a brilliant idea, I like it! And indeed it makes sense to keep the queue even between restarts.
Adds an option to make a persistent interceptor queue, so messages won't be lost if connection is interrupted.