I am working on a proof of concept implementation of NServiceBus v4.x for work.
Right now I have two subscribers and a single publisher.
The publisher can publish over 500 message per second. It runs great.
Subscriber A runs without distributors/workers. It is a single process.
Subscriber B runs with a single distributor powering N number of workers.
In my test I hit an endpoint that creates and publishes 100,000 messages. I do this publish with the subscribers off line.
Subscriber A processes a steady 100 messages per second.
Subscriber B with 2+ workers (same result with 2, 3, or 4) struggles to top 50 messages per second gross across all workers.
It seems in my scenario that the workers (which I ramped up to 40 threads per worker) are waiting around for the distributor to give them work.
Am I missing something possibly that is causing the distributor to be throttled? All Buses are running an unlimited Dev license.
System Information:
Intel Core i5 M520 # 2.40 GHz
8 GBs of RAM
SSD Hard Drive
UPDATE 08/06/2013: I finished deploying the system to a set of servers. I am experiencing the same results. Every server with a worker that I add decreases the performance of the subscriber.
Subscriber B has a distributor on one server and two additional servers for workers. With Subscriber B and one server with an active worker I am experiencing ~80 messages/events per second. Adding in another worker on an additional physical machine decreases that to ~50 messages per second. Also, these are "dummy messages". No logic actually happens in the handlers other than a log of the message through log4net. Turning off the logging doesn't increase performance.
Suggestions?
If you're scaling out with NServiceBus master/worker nodes on one server, then trying to measure performance is meaningless. One process with multiple threads will always do better than a distributor and multiple worker nodes on the same machine because the distributor will become a bottleneck while everything is competing for the same compute resources.
If the workers are moved to separate servers, it becomes a completely different story. The distributor is very efficient at doling out messages if that's the only thing happening on the server.
Give it a try with multiple servers and see what happens.
Rather than have a dummy handler that does nothing, can you simulate actual processing by adding in some sleep time, say 5 seconds. And then compare the results of having a subscriber and through the distributor?
Scaling out (with or without a distributor) is only useful for where the work being done by a single machine takes time and therefore more computing resources helps.
To help with this, monitor the CriticalTime performance counter on the endpoint and when you have the need, add in the distributor.
Scaling out using the distributor when needed is made easy by not having to change code, just starting the same endpoint in distributor and worker profiles.
The whole chain is transactional. You are paying heavy for this. Increasing the workload across machines will really not increase performance when you do not have very fast disk storage with write through caching to speed up transactional writes.
When you have your poc scaled out to several servers just try to mark a messages as 'Express' which does not do transactional writes in the queue and disable MSDTC on the bus instance to see what kind of performance is possible without transactions. This is not really usable for production unless you know where this is not mandatory or what is capable when you have a architecture which does not require DTC.
Related
On our RabbitMQ installed in production, we have a performance issue.
To explain the context, we have an initialization batch that creates around ~60k messages. For business reasons, those messages must be treated in strict order and we can't lose any. As such, we have only one queue which is durable and lazy and one consumer (SpringBoot AMQP) with a prefetch of 10. Both are on the same virtual machine.
At first, the processing is fast enough, around 5 to 10 messages per second. But it progressively slows down until it reaches a cap of fewer than 20 messages per hour. It takes approximately 1 hour to reach this point.
After some investigations, we found out that the problem comes from RabbitMQ. When we simply stop and restart it, the performance goes back to normal and then drops slowly again. Doing the same on just the consumer doesn't change anything.
I'm thinking about some resources bottleneck but I can't manage to find which one as RAM, CPU, and disk looks fine. I am not really familiar with ERL virtual machine and managing RabbitMQ itself so I may have missed something.
Does someone as an idea of the source of the problem or where I could look for more information on what is happening?
RabbitMQ characteristics :
ERL 23.3.2
RabbitMQ 3.8.14
I'm using RabbitMQ 3.6.10.
Having 16GB RAM on the machine and set water benchmark to 6GB. 4 cores.
I'm trying to perform some tests on Rabbit. Creating 1 publisher and no one that will consume the messages.
When creating 1 connection with 1 channel publishing unlimited messages one after another the management UI shows that average publish/s in ~4500.
When increasing the number of channels/connections and do it parallel in different kinds of combination i can see that it also not writing more than ~4,500.
I saw many benchmarks that talk about many more messages per second.
I can't figure what can cause the bottleneck? Any ideas?
In addition, when using many channels with many messages I get to some point that the Rabbit RAM is full and it blocks the publishers from publishing more messages. This is a good behavior but the problem is that the Rabbit stops writing to the disk and it stuck in this status forever. Any ideas?
In current project we currently use 8 worker role machines side by side that actually work a little different than azure may expect it.
Short outline of the system:
each worker start up to 8 processes that actually connect to cloud queue and processes messages
each process accesses three different cloud queues for collecting messages for different purposes (delta recognition, backup, metadata)
each message leads to a WCF call to an ERP system to gather information and finally add retreived response in an ReDis cache
this approach has been chosen over many smaller machines due to costs and performance. While 24 one-core machines would perform by 400 calls/s to the ERP system, 8 four-core machines with 8 processes do over 800 calls/s.
Now to the question: when even increasing the count of machines to increase performance to 1200 calls/s, we experienced outages of Cloud Queue. In same moment of time, 80% of the machines' processes don't process messages anymore.
Here we have two problems:
Remote debugging is not possible for these processes, but it was possible to use dile to get some information out.
We use GetMessages method of Cloud Queue to get up to 4 messages from queue. Cloud Queue always answers with 0 messages. Reconnect the cloud queue does not help.
Restarting workers does help, but shortly lead to same problem.
Are we hitting the natural end of scalability of Cloud Queue and should switch to Service Bus?
Update:
I have not been able to fully understand the problem, I described it in the natual borders of Cloud Queue.
To summarize:
Count of TCP connections have been impressive. Actually too impressive (multiple hundreds)
Going back to original memory size let the system operate normally again
In my experience I have been able to get better raw performance out of Azure Cloud Queues than service bus, but Service Bus has better enterprise features (reliable, topics, etc). Azure Cloud Queue should process up to 2K/second per queue.
https://azure.microsoft.com/en-us/documentation/articles/storage-scalability-targets/
You can also try partitioning to multiple queues if there is some natural partition key.
Make sure that your process don't have some sort of thread deadlock that is the real culprit. You can test this by connecting to the queue when it appears hung and trying to pull messages from the queue. If that works it is your process, not the queue.
Also take a look at this to setup some other monitors:
https://azure.microsoft.com/en-us/documentation/articles/storage-monitor-storage-account/
It took some time to solve this issue:
First a summarization of the usage of the storage account:
We used the blob storage once a day pretty heavily.
The "normal" diagonistics that Azure provides out of the box also used the same storage account.
Some controlling processes used small tables to store and read information once an hour for ca. 20 minutes
There may be up to 800 calls/s that try to increase a number to count calls to an ERP system.
When recognizing that the storage account is put under heavy load we split it up.
Now there are three physical storage accounts heaving 2 queues.
The original one still keeps up to 800/s calls for increasing counters
Diagnositics are still on the original one
Controlling information has been also moved
The system runs now for 2 weeks, working like a charm. There are several things we learned from that:
No, the infrastructure is "not just there" and it doesn't scale endlessly.
Even if we thought we didn't use "that much" summarized we used quite heavily and uncontrolled.
There is no "best practices" anywhere in the net that tells the complete story. Esp. when start working with the storage account a guide from MS would be quite helpful
Exception handling in storage is quite bad. Even if the storage account is overused, I would expect some kind of exception and not just returning zero message without any surrounding information
Read complete story here: natural borders of cloud storage scalability
UPDATE:
The scalability has a lot of influences. You may are interested in Azure Service Bus: Massive count of listeners and senders to be aware of some more pitfalls.
I have a publisher pushing to a queue at a slightly larger rate than the consumers can consume. For small numbers, it is okay, but for a very large number of messages, RabbitMQ starts writing it to the disk. At a certain point of time, the disk becomes full, and flow control is triggered. From then on, the rates are really slow. Is there any way to decrease or share this load between cluster nodes? How should I design my application so that flow control is never triggered?
I am using RabbitMQ 3.2.3 on three nodes with 13G RAM, and 10G of system disk space - connected to each other through the cluster. Two of these are RAM nodes, and the remaining one is a disk node, also used for RabbitMQ management plugin.
You can tweak the configuration, upgrade hardware etc and in the end you'd probably want to put a load balancer in front of your RabbitMQ servers to balance the load between multiple RabbitMQ nodes. The problem here is that if you are publishing at a higher rate than you are consuming, eventually you will run into this problem again, and again.
I think the best way to prevent this from happening is to implement logic on the publisher side that keeps track of the number of requests waiting to be processed in the queue. If the number of requests exceeds X the publisher should either wait until the number of messages has gone down, or publish new messages at a slower rate. This type of solution of course depends on where the messages published are coming from, if they are user submitted (e.g. through a browser or client) you could show a loading-bar when the queue builds-up.
Ideally though you should focus on making the processing on the consumer side faster, and maybe scale that part up, but having something to throttle the publisher when it gets busy should help prevent buildups.
I'm using 6 servers to make a cluster and they are all disk nodes. I use rabbitmq for collecting log file for our website. Now at the peak hour, the publish rate is about 30k message per second. There are 2 main consumers(hdfs and elasticsearch) and each one need to handle all message, so the delivery rate hit about 60k per second.
In my scenario, a single server can hold 10k delivery rate and I use 6 node to load balance the pressure. My solution is that I created 2 queues on each node. Each message is with a random routing-key(something like message.0, message.1, etc) to distribute the pressure to every node.
What confused me is:
All message send to one node. Should I use a HA Proxy to load balance this publish pressure?
Is there any performance difference between Durable Queues and Transient Queues?
Is there any performance difference between Memory Node and Disk Node? What I know is the difference between memory node and disk node is only about the meta data such as queue configuration.
How can I imrove the performance in publish and delivery codes? I've researched and I know several methods:
disable the confirm mechanism(in publish codes?)
enable HiPE(I've done that and it helped a lot)
For example, input is 1w mps(message per second), there are two consumers to consume all message. Then the output is 2w mps. If my server can handle 1w mps, I need two server to handle the 2w-mps-pressure. Now a new consumer need to consume all message, too. As a result, output hits 3w mps, so I need another one more server. For a conclusion, one more consumer for all message, one more server?
"All message send to one node. Should I use a HA Proxy to load balance this publish pressure?"
This article outlines a number of designs aimed at distributing load in RabbitMQ.
"Is there any performance difference between Durable Queues and Transient Queues?"
Yes, Durable Queues are backed up to disk so that they can be reinstated on server-restart, for example. This adds a nominal overhead, though the actual process occurs asynchronously.
"Is there any performance difference between Memory Node and Disk Node?"
Not that I'm aware of, but that would depend on the machine itself.
"How can I imrove the performance in publish and delivery codes?"
Try this out.