Ansible on CentOS 10

We mostly utilize Ansible  for deployment automation and CI/CD pipelines.

I have used some of the Ansible  libraries for some of the deployments. The way conditions are handled in Ansible, such as skip conditions or failure conditions, can be complex with multiple conditions, but there is support for using them. 

Additionally, the automation capabilities streamline deployment processes, providing reliability and reducing manual intervention and errors.

More library support for microservices architecture and Kubernetes  would be helpful. There is also a need to improve the handling of conditions, which can be tricky and require the use of flags.

I have been working with Ansible for about six to seven years.

Ansible is quite stable. There are infrastructure-wise reliability and fewer issues, although network issues might cause some failures.

Ansible can face scalability issues, such as limitations when trying to scale up infrastructure. It might struggle with connection dropping or spawning additional VMs under certain conditions.

I have used both Ansible and Terraform . Ansible can lag when compared to Terraform  for certain microservice deployments.

The initial setup can be challenging and is not very intuitive. A person needs to be knowledgeable in Ansible, and it should be well documented.

The benefits include the maintainability of the existing environment, especially a hybrid infrastructure. Ansible makes scripting and learning processes better and easier.

There is a pricing associated with Ansible, however, I find it reasonable.

I would recommend Ansible. That said, it depends on the infrastructure and whether an off-premises or on-premises cloud is used. 

I would rate Ansible between eight and nine.

We use Red Hat Ansible Automation Platform  for the automation of our servers and applications. We also use both Terraform  and Ansible  to automate our infrastructure.

Red Hat Ansible Automation Platform has saved our time and reduced errors in our processes. It used to take us two months to provide a server in our organization, and now it only takes a few minutes to have the same server. Automation has ensured that tasks are done in the same way every time, reducing the likelihood of errors.

The capacity to install products on the operating system is very valuable. Ansible is better at handling the final configuration of servers. We prefer Terraform  for creating multiple resources in a project, but Ansible is better for final configurations.

Ansible's centric idea of servers needs to be changed. In modern infrastructure, there are more than just servers. The initial server-centric approach in Ansible is a bit strange. It should improve its functionality with cloud resources like Azure , AWS , or Google Cloud . 

Ansible could also improve its capabilities in managing several resources at the same time, similar to Terraform. Moreover, more integration with other tools would be beneficial.

I have been using Red Hat Ansible Automation Platform for about three to four years.

The stability of Ansible is rated high, around eight to nine on a scale of ten.

We haven't experienced any problems with Ansible's scalability. We use it at the project level and create around ten to 20 resources. We haven't tested it with thousands of servers, so it's difficult to say how it would perform in such scenarios.

We are using the free version of Ansible, and so far, the support has been very high, considering that it is a free version. We are in discussions with Red Hat and IBM about possibly purchasing the commercial version when we start using Ansible for patching servers.

Positive

Before Ansible, we used BladeLogic from BMC. We switched to Ansible as it was easier to use, had more functionality, and there were more people in the market who knew Ansible compared to BladeLogic. Overall, Ansible is a much better product.

The setup of Ansible is easy. It's faster to start working with Terraform. However. Ansible's setup is also straightforward. The basic installation process is quick and effortless.

I recommend using both Ansible and Terraform for automation, especially now that both are under IBM.

I'd rate the solution eight out of ten.

We are primarily using Ansible  for automation purposes as it is a configuration management tool. It is utilized for various activities such as DNS activity, changes to web servers, virtual host settings, and other day-to-day tasks, all of which are templated in Ansible .

Ansible allows us to manage a multitude of servers efficiently. We can deploy configurations and changes effectively and gather detailed reports. This means we have substantial control and flexibility in managing our servers.

I can do anything with Ansible. It allows control over thousands of servers, whether virtual or physical. The flexibility to manage deployments, configuration changes, and reporting is highly valuable. Ansible is containerized, making it easy to pull updated containers for automation.

There are challenges in using the graphical interface, particularly in open-source versions. The Subscription model presents some limitations, and there is room for improvement in making the Ansible navigator more flexible for open-source use. Installation can also be challenging, especially for graphical components.

I have been working with Ansible since 2012.

Ansible is very stable. There are no issues concerning the system's stability when managed with Ansible.

Ansible provides fantastic scalability. This tool allows us to manage a significant number of clients without limitations, making it suitable for large-scale operations.

I rate Red Hat's customer support for Ansible at nine points out of ten. Customer support for Ansible is excellent, and any issues we have encountered have been resolved promptly.

Positive

Setting up Ansible is relatively straightforward. Installing the core product takes about thirty minutes to an hour. However, fully setting up Ansible with additional servers might take around two to three hours.

The implementation is handled by myself and one other colleague.

There is a need for more flexibility in the subscription model, but I do not have detailed insights into the pricing and licensing setup.

I'd rate the solution nine out of ten.

Our use cases are related to system compliance, package management, system patching, and general infrastructure and security as a code.

We can easily remediate a lot of the routine tasks or even break-fix issues that keep coming up. It could be a simple break for which we do not have a root cause analysis for a proper solution, but it keeps coming up. We can easily remediate such issues and have those not impact workflows for our developers while we work on a full solution in the backend. It also helps with the general tasks of maintaining systems or correcting them when things get out of line. We are able to deploy things at scale versus having someone there. Their time is much more valuable. They do not have to spend five or ten minutes here or there throughout the day when they can be focused on a more dedicated effort elsewhere.

It helps us achieve our mission because we want to enable our researchers and developers to have an ecosystem that allows them to do what they need but in a secure environment. We want to protect system information, data, and IPs. We want to be able to do that at scale, quickly roll out changes, and maintain systems as needed for compliance. We want to allow all that holistically for our groups to manage. It has helped us with workflows to help expedite that process.

We also use other Red Hat products. We use Satellite and IDM, and we are moving towards AAP and EDA. We chose these products to help manage our infrastructure. Because we are managing Linux systems, we want everything to work and play with each other as well as they can. We are adopting the same ecosystem. Choosing Red Hat products allows us to do that with the most reliability, and we also get support from the experts who built the programs themselves. If anything were to go wrong, we have that backup of people who are dedicated to helping us find the solution for it. With multiple Red Hat products, we have interoperability. We are able to spread across different teams. If we standardize our process on the same products and the same technologies, it means we can train other departments and other groups to manage their own piece of the pie. If they have many different products, the only people who would know how to work with a product are the ones who are dedicated to working with that product, whereas if we standardize on the same kind of ecosystem, it is easier to cross-train on that.

It does a good job because the whole syntax is human-readable, and it is YAML. While it has its quirks, it is pretty straightforward. Generating playbooks, tasks, modules, and other things is pretty easy. Even the plugins are Python-based, so it is an easy language and it is near readable for Python, whereas, trying to make scripts for something or code in different languages is very tedious. You need people who are experts in making that deployable and handling any unexpected outcomes or different conditions that may cause issues, so having something that is easy and straightforward to learn and that handles all of the guesswork on the backend makes it much easier to develop automation.

It helps reduce the training required to learn how to automate things. It is a very easy language to learn, and it is not like learning an entire programming language.

It helps connect teams, such as developers, operations, or security so that they can automate together. Developers and operations definitely work hand in hand. Introducing infrastructure as a code or security as a code when they have not adopted that mindset can be difficult. They definitely agree with the concept but getting them to work and manage their own playbooks is a little bit more difficult because not everyone is intrinsically interested in coding even if it is a simple language. However, being able to have the partnership to say that we can help them out and they have a readable syntax is helpful. They can clearly see what is going on. It helps them a lot in understanding what is going on in the backend. We are becoming more and more productive as we get more products involved and more people on board with developing and adopting their own respective departments as code, such as infrastructure, security, IT, and anything else that we may have in the pipeline.

Its collection of certified content helps out a lot. We are able to pull certified content and then use it within whatever we are trying to deploy. Everything essentially works with whatever we are trying to deploy. If it does not work, we know that it is something on our end. It helps out with identifying known goods and then working from there.

It helps to reduce the time we spend on low-value or repetitive tasks. It has taken off about 45% of my workload from tedious tasks.

The easy-to-read syntax for YAML files and the interoperability between modules are valuable. There is also the ability to develop your own plug-ins and modules to be able to contribute. The scalability of Ansible across different environments is also valuable.

There are some integration issues with other technologies such as Satellite. It could be a Satellite issue and not an Ansible issue. There are bugs related to failed tasks not being failed and then reverting back to completed tasks. It might not be because of Ansible. It could be because of Satellite.

There have been some differences between the operating systems that we have noticed. It could be down to cryptographic policies, but we have noticed some speed issues. They should work on the speed of deployment on different operating systems.

They have already managed the other issue of introducing execution environments to make sure everything is the same with every system.

We have been using this solution for about two years.

It has been pretty reliable. The only issues that have come up are because of people asking it to do things that it is not built to do. All our issues tend to be someone on our end doing something that they probably should not be doing. If you are using it as it is intended, it is pretty reliable.

It is scalable. We have plans to increase its usage. We are planning to implement AAP, EDA, etc.

The support has been nice whenever we had to put in any Red Hat-related issues. They are fairly responsive.

If there are issues that exceed their knowledge base, they usually escalate it to someone else who can handle them. We usually get a working solution fairly quickly with an actual root cause diagnosis.

I would rate them a ten out of ten. They have always been pretty speedy and reliable.

Positive

We used to use Puppet. It is not the same. It is agent-based, and I am not a fan of it. Ansible is much more lightweight, robust, and flexible, so our other alternative used to be Puppet, but we are now onboard with Ansible.

In terms of ease of use, Ansible is a lot easier to read and develop with. Red Hat has certified content. They have their own supported modules, and the community also updates the modules that are community-maintained, whereas Puppet does not have the same kind of development. You are waiting for things to be updated from Puppet's end, which does not quite work with our use case. If there is a critical need for the update of a certain module to work with our environment, it does not really work because we need things to be updated semi-frequently or at least be maintained. Ansible helps take care of that. With Ansible, we can work with push or pull configurations depending on what we want to set up instead of having one configuration that Puppet works with, so it gives us flexibility.

Ansible is cheaper than Puppet, but we are scaling out and flushing out more infrastructure. So, because of how we use it, it does associate more costs. We use different kinds of licensing and products, but it is worth the costs and investments, whereas Puppet is its own ecosystem. It does its own thing, and that is it. It does not scale out to do the kind of things that we want to do.

We manage from a central location and push out to our remote sites. Everything is on-prem.

It was simple from the Ansible's end. The complicated thing was getting the other tools and architecture needed to push out to where we wanted and getting all that developed, but the initial push out with Ansible was pretty simple and straightforward.

Our deployment strategy was to get it running and push it out to showcase the proof of concept of how much time we spend doing this and how much time we can save now because it can handle that automatically. The game plan was not a true holistic view of what we could do moving forward because we could not do that until we got more buy-in from everyone else to show how this positively impacts the organization.

We did it on our own. We got the licensing and bought the tools. It was pretty much on us to figure it out.

The scalability of it is the biggest return on investment because it can scale to a small handful or thousands. It works for simple and complex scenarios. It can all be done without a lot of backend research. Obviously, with more complexity, you need people who know a bit more, but it is pretty easy to get up to speed. It is pretty flexible in terms of how it can scale to different environments.

Everything is generally fair. No one ever likes to pay a lot of money, but we are getting the value. We also get support with it. It has been fair and worthwhile.

We did not evaluate any other solution. It was pretty much trying to buy more into Puppet, but as we determined needs, we scaled it out to Ansible because we work in an environment with Red Hat. We also have other operational needs for which it works.

I would rate it a ten out of ten because I enjoy it a lot.