HAProxy is very weak in the logging and monitoring part and requires improvement.

In the future, the tool should have additional modules for different purposes.

There is room for improvement in the pricing model. It could be cheaper. 

The solution can be improved by controlling TCP behavior better rather than just reporting them in the logs. Automatic blocking of clients that don't properly close their TCP session (CH/CD) would be good.

Although HAProxy is essentially open-source, many features are not available.

While troubleshooting, we are having some difficulties. There are no issues when it is running; it is stable and very good; however, if there is a troubleshooting issue or an incident occurs, we will have issues because this is open-source.

We are only experiencing problems at that time. Otherwise, everything is fine.

Because we need to search for a document or some troubleshooting information. That could be beneficial to us.

To resolve the troubleshooting issues, additional documentation and troubleshooting are required.

The product is satisfactory. However, additional documentation, additional technical documents, and troubleshooting steps are the types of things that can only make this solution better. Also, more clarity on where the package is hitting and where it is stopping is needed.

HAProxy could do with some good combination integrations. 

The product should have more security and dashboard functionality for monitoring so that any administrator can see the usability and track all the incoming and outgoing requests. It should have a better dashboard GUI, and more security models should be there. The GUI should be more responsive and show the detailed output of logs. We need it to be more visible.

Documentation could be better.

The reconfigurability in terms of the tooling could be improved and maybe an editor plugin can be added. Something that might help with syntax or common use cases.

There is room for improvement in HAProxy's dynamic configuration. Currently, dynamic changes are lost when reloading the service, and it would be beneficial if dynamic configuration changes could be applied without losing the configuration or reloading the service, ensuring backups and preserving the static configuration.

HAProxy could improve by making the dashboards easier to use, and better reports and administration tickets.

The logging is pretty hard to understand, but the documentation for the logging is decent. That would be my only criticism. Sometimes it's challenging to get through the log, and you need a log to understand what is going on. It isn't easy to map the logging with the documentation, and every time I read the log, I have to pull out the documentation to understand what I'm reading.

And there is some more functionality that I would like to see. For example, you'll do a TLS to the front leg— whatever connects to your load balancer. You do the HTBS or whatever TLS connection there. And then, on the back end, you usually have to clear it a lot of times. I want to be able to do TLS all the way through on both legs. I don't know if it can do that. HAProxy might be able to do this already, but I haven't done enough research to see if this is possible

I'd like to see better documentation and preferably a French version as well. The product is used a lot here and that would be helpful.

The web stats UI, which provides the status of the health and numbers, could greatly benefit from having a RESTful interface to control the load-balanced nodes. Although there is a hack around the UI (by issuing a POST request to HAProxy with parameters), a RESTful interface would greatly improve the automation process (through Chef and Ansible).

Dynamic update API. More things should be possible to be configured during runtime.

We would like to see dynamic ACL and port update support. Our infrastructure relies on randomly allocated ports and this feature would allow us to update without restarting the process. The ACL add/update would help with some direct routing challenges that currently require us to work around them with a map and static back-ends.

HAProxy running in multiple cores, for example one for HTTP and another for HTTPS, requires the use of "nbproc". So if nbproc = 2, you will have two processes of HAProxy running. However, the stats of HAProxy are not aggregated, meaning you don't really know the collective status in a single point of view. Each process has its own socket and it's up to you to aggregate them, and then your stats become less accurate.

Also, having multiple HAProxy nodes in High Availability mode requires the use of clustering software such as Pacemaker and Corosync which are very complex.

There is no standardized document available. So, any individual has to work from scratch to work it out. If some standard deployment details are available, it would be helpful for people while deploying it. There should be more documentation on the standard deployment.

When you have to customize it for your application requirements, there are a lot of challenges. There should be more support for customization. To customize it better, there should be some kind of programming integration.

The logging functionality could use improvement, as it is a little cryptic. Additional logging functionality with better documentation would be helpful.

The documentation is extremely detailed, but I think it could be restructured so it isn't so daunting for beginners.

The product has the basic foundations of any load-balancing product. It is simple and basic. The product does not have any new technologies.

The basic clustering is not usable in our very specific setup. The clustering is mainly a configuration replication and is great in a case of active-passive usage. In the case of an active-active (or with more than two nodes) where the configuration is not fully identical, it cannot be used as-is.

The only area that I can see needing improvement is the management interface, since it is pretty much all through the CLI or configuration. However, HAProxy does have another product that we evaluated called ALOHA, which has a web front-end, but we found it did not meet our needs.

The visibility could be improved. Stability can falter when there is a lot of traffic. End-to-end connections are limited. 

I would like to see better search handling, and a user interface, with a complete functional graphical unit.

The GUI is clunky and the on-board metric viewing solution is poor. We ship metrics off the appliance to InfluxDB (using CollectIT, which is bundled with the appliance), to view them directly with Grafana.

The solution has bad performance issues. 

A better GUI would be nice.

We've changed solutions as it doesn't fit with our current needs. 

While the product is quite perfect, it could use more supervision and be more active. Also, more simplicity, more peripherals, and more scalability.

I would improve the web-based UI, but it’s a matter of personal preference.

We need to handle new connections by dropping, or queuing them while the HAProxy restarts, and because HAProxy does not handle split config files.

In my opinion, there are three main areas to improve:

1. Make remote management more modern by adding API.
2. Propose a general HA solution for HAProxy (no I'm using keepalived for this).
3. Thread option should be a bit more stable.

From my point of view, the configuration should be more friendly, perhaps with a Web interface. For example, I work with the ClusterControl product for Severalnines, and we have a Web interface to deploy the HAProxy load-balancer.

They should introduce one feature that I know many people, including me, are waiting for: HAProxy should have provide hot-swipe for back-end servers, no downtime while shifting back-end servers.

They also need a more detailed GUI for monitoring and configuration.

It needs proper HTTP/2 support.

I would like to evaluate load-balancing algorithms other than round robin and SSL offloading.

Also, it would be helpful if I could logically divide the HAProxy load-balancing into multiple entities so that I would install one HA Proxy LB application which could be used for different Web servers for different applications. I am not sure if these features are available.

Maybe HAProxy could be more modular. 

• Availability of REST API for Configuration and Reporting
• An even bigger suite of Load Balancing algorithms
• Documentation

I would like to see a two (or more) native cluster support without third parties or DNS manipulations.

HAProxy needs to improve its ability to handle a lot of connections, to be more like NGINX.

Pricing, monitoring, and reports can be improved.