Management UI: extend Get Message feature to streams#11030
Management UI: extend Get Message feature to streams#11030LoisSotoLopez wants to merge 1 commit intorabbitmq:mainfrom
Conversation
|
|
||
| -module(rabbit_mgmt_wm_queue_get). | ||
|
|
||
| -include_lib("kernel/include/logger.hrl"). |
There was a problem hiding this comment.
is this used anywhere? or just a leftover from debugging?
| ReqData, Context) | ||
| end) | ||
| end). | ||
| Resource = rabbit_misc:r(<<"/">>, queue, Q), |
There was a problem hiding this comment.
| Resource = rabbit_misc:r(<<"/">>, queue, Q), | |
| Resource = rabbit_misc:r(VHost, queue, Q), |
| {redelivered, Redelivered}, | ||
| {exchange, Exchange}, | ||
| {routing_key, RoutingKey}, | ||
| {consumer_tag, ConsumerTag}, |
There was a problem hiding this comment.
consumer tag is just hardcoded "ctag" so it does not need to be included.
| delivery_tag = DeliveryTag, | ||
| consumer_tag = ConsumerTag}, | ||
| #amqp_msg{props = Props, payload = Payload}} -> | ||
| ok = amqp_channel:cast(Ch, #'basic.ack'{delivery_tag = DeliveryTag, |
There was a problem hiding this comment.
I dont think we want to ack the message. If the stream has 20 messages, and we ask for 10 (prefetch = 10) when we ack the first message, the broker will send the 11th message. (Prefetch count means the max number of outstanding ie unacknowledged messages)
There was a problem hiding this comment.
This feature should not arguably use AMQP 0-9-1 at all. And the original feature should never have used an AMQP 0-9-1 Erlang client to begin with.
There was a problem hiding this comment.
Yes it would probably be better to use rabbit_queue_type directly to get a message off the stream.
There was a problem hiding this comment.
I dont say that an amqp connection should be used to implement this feature, just want to mention the aspect, or implicit benefit that the server side of an amqp connection does additional authorization checks.
Would like to cite the recent attempt for the delete queue endpoint to use an internal API #9550 which was reverted #10062 (comment). Maybe can you recall what was the difficulty, why the path outlined in #10062 was abandoned?
Or maybe that restriction is not applicable for a simple reading from a queue?
There was a problem hiding this comment.
Sounds like a way to not move anywhere.. why would not using the 0-9-1 protocol be ok expect it's maybe not 100% optimal? This solves a real world problem with minimal change.
| amqp_channel:subscribe( | ||
| Ch, | ||
| #'basic.consume'{queue = Queue, | ||
| no_ack = NoAck, |
There was a problem hiding this comment.
NoAck is always false so we can hardcode it I think
Also hardcode the CTag
|
Before you spend any more time on this feature, have you considered consulting with the core team as to whether this feature has any chance of being accepted? Or, perhaps, whether it should use the Erlang AMQP 0-9-1 client? Or any client at all and not an internal API? I'd very much like to see contributors, even those who contribute mere 1-2% of R&D despite directly making money off of RabbitMQ, to spend their time on more fundamental issues like #11038 or #10763 (two examples of much more fundamental problems it does not take that deep an understanding of RabbitMQ internals to work on). Specifically around streams, rabbitmq/rabbitmq-website#1885 is infinitely more important for stream adoption than this feature. |
michaelklishin
changed the title
5 participants
Proposed Changes
Extends the management endpoint for retrieving messages from a queue allowing this operation to be applied to stream queues. Alters the
queue.ejstemplate to show the Get messages box when showing a stream queue information.Types of Changes
What types of changes does your code introduce to this project?
Put an
xin the boxes that applyChecklist
Put an
xin the boxes that apply.You can also fill these out after creating the PR.
If you're unsure about any of them, don't hesitate to ask on the mailing list.
We're here to help!
This is simply a reminder of what we are going to look for before merging your code.
CONTRIBUTING.mddocumentFurther Comments