AWS CodePipeline Reviews

I use CodePipeline with CodeCommit, CodeBuild, and CodeDeploy.

My main use case is around building a small product. We've created a framework by leveraging AWS and Azure services. It functions as a table-driven approach, it allows us to switch features of the product on or off.

We've set up CI/CD pipelines to enable these services, ensuring security and automatically propagating framework tables and metadata into the target environment.

For example, as part of a customer's day zero setup, we provide them with the necessary tools to start their development. We utilize AWS CodeCommit to store our product components and then enable the CI/CD pipeline to seamlessly transfer these components from our tenancy to the customers.

The functionality allows you to define multiple steps and dynamically create or destroy agents as required, saving costs. You can also connect to various repositories like AWS CodeCommit, GitLab, or GitHub. 

Moreover, you can trigger other jobs or even other pipelines from one CI/CD pipeline.

While I haven't personally tested it, the documentation suggests that it's possible to call Azure Repos or Jenkins pipelines/jobs from AWS CodePipeline.

From my perspective, some in-built capabilities could be enhanced. For example, SMS notifications should be integrated directly into the service. Currently, if I want to send an email notification when a pipeline or service fails, I need to configure it separately and hook it in.

A couple of similar functionalities seem to be embedded within Jenkins CI/CD pipelines. 

Additionally, a more comprehensive UI would be beneficial, one that visually represents the progress of the pipeline execution. If a pipeline has ten steps, I should be able to see the execution in real-time as each step progresses.

So, I'm expecting that an AWS monitoring-like control where I can track which steps are being executed, view the populated logs, and even have email notifications sent to stakeholders or agents as needed. These out-of-the-box capabilities would significantly streamline the development process.

So, I would like to see in-built capabilities to populate the logs and some kind of log analysis to determine whether the job failed or passed. Based on that, notifications should be sent to other stakeholders.

I have been using it for one year. 

I would rate the stability a ten out of ten.

I would rate the scalability an eight out of ten. So there are no issues with scalability, it's pretty scalable.

One of my solutions was reviewed by an AWS architect. Based on their support and guidance, I would rate it nine out of ten.

Positive

I work with AWS Glue, AWS Terraform, AWS CodePipeline, and AWS Lambda. I have also worked with S3 buckets.

For Azure, I use Azure Databricks and Azure Data Factory.

The initial setup is pretty simple. The documentation is well-defined, and a lot of help and support is available within AWS.

The solution is deployed in the AWS cloud. It's a cloud solution.

Deployment time: 

As we do some metadata ingestion, framework tables are created, and DDLs and DMLs are executed. So, sometimes, it takes 10 to 15 minutes to complete the entire framework setup. We have multiple pipelines. 

1. The first pipeline provisions the environments and enables DB connections between different services. 
2. The second executes DDLs and DMLs in the database. 
3. The third pipeline pulls existing Glue jobs used for data transmission. 

In a nutshell, everything is completed within ten to fifteen minutes, and the environment is usable.

The core benefit is that it saves developers' time. It's a great choice for agile development for any of the cloud vendor, it enables CI/CD pipelines so developers can test quickly. 

If a developer commits code by 4 PM, the queue can trigger the build and, if auto-build and auto-deploy are enabled, automatically build and deploy. Otherwise, the queue can build and deploy it, and someone can do a sanity check. In a single workday, features can be tested and enabled.

Another big advantage is observability. I can monitor the logs, put additional logs, and check the status of my steps and jobs. Overall, the pipeline helps accelerate development, monitoring, and observability, and it satisfies the non-functional requirements of reliability.

The pricing is manageable. AWS and Azure have a pay-as-you-go model. So the price is not a big factor. It's manageable, not that high yet.

I can compare it with Azure DevOps, and I can say it's a comparison between Jira and GitHub. Jira, GitHub, and TeamCity also allow us to develop CI/CD pipelines by integrating with Jira, Confluence, and GitHub. Jenkins is also there, which offers similar features. I used Jenkins five years ago.

I would recommend it for at least medium, but especially enterprise level. They should go for AWS or Azure. Both are really fine and good.

I currently work with four different customers as a technical architect for data migration and a couple of other projects. My recommendation is to go for AWS CI/CD and leverage AWS services.

Overall, I would rate it a nine out of ten. I can recommend it to others. 

Our primary use case for CodePipeline involves deploying the different services, such as hosting a website on ECS or EC2 and deploying source code on container services or EC2 instances. We configure the source code with remote repositories like GitHub or Bitbucket, build the code, and store images in ECR. We use these images for deployments in ECS clusters.

CodePipeline has helped our organization by simplifying the deployment of services for small and medium customers on EC2 or ECS clusters. Customers have saved time with this tool.

CodePipeline allows integration with any source code and facilitates deployment to EC2 or ECS, which is highly valuable. For software delivery, it provides excellent integration capabilities with AWS services.

While CodePipeline's UI has improved, further enhancements in user experience are necessary to make it more intuitive. Adding new features would allow for easier management by anyone.

We have been working with CodePipeline for two years.

CodePipeline is highly stable, with 100% uptime and no failures experienced.

CodePipeline offers excellent scalability, allowing swift and adaptable deployments based on customer requirements.

I have not contacted AWS support for CodePipeline, so I cannot provide a detailed review about their customer service.

Positive

The initial setup was straightforward once familiar with the process. It took about half an hour for the first deployment.

Customers have benefited from CodePipeline by saving time and resources, though detailed metrics on ROI were not provided.

CodePipeline's pricing is affordable and average overall. I would rate it a two on a scale where ten is the highest price.

CodePipeline is a very good AWS service suitable for small, medium, and enterprise customers, depending on the use case.

I'd rate the solution nine out of ten.

I am a freelancer, and I usually work with many companies across the globe. I have been doing it for over the last five years. I work with various AWS services, mainly focusing on ECS, EC2, S3, CodePipeline, and RDS. 

Additionally, I work with serverless services like Lambda, API Gateway, Cognito, and DynamoDB. I use CodePipeline for continuous integration and continuous deployment, deploying to Elastic Beanstalk and ECS. I also use CodeBuild to build Docker images and use CodeDeploy for ECS clusters. I deploy CloudFormation stacks using CodePipeline for infrastructure provisioning and deploying to Lambda.

Recently, we deployed our microservices application into ECS cluster, leveraging CodePipeline for deployment across different environments like dev, staging, and prod. CodePipeline helps with unit testing using CodeBuild and deploying directly to ECS. It's robust and reliable for our deployment process.

I prefer using CodePipeline and CodeBuild in AWS due to their integration with AWS services, like directly deploying to ECS using Cloud Deploy. CodePipeline allows for canary deployments, blue and green deployments out of the box, making deployments robust and easier to roll back.

I would appreciate if we don't have to install any agents, even for EC2, to deploy using CodeDeploy into EC2 servers. It would help if AWS allows application deployment without requiring the Fortify agent installation.

I have been using CodePipeline for around five years.

Stability is quite good; I would rate it nine or ten. There are no significant stability issues, although sometimes CodeBuild can be slow.

Scalability is excellent, I would rate it ten out of ten. We use it extensively for deployments and have not faced any issues.

Customer support is better with a business plan. Basic and developer plans are not as responsive, sometimes taking two to three days to respond, even for important issues.

Neutral

I would say it is cheaper than other solutions because I use many others.

AWS charges based on builds, not too expensive compared to alternatives.

I encourage customers to use CodePipeline due to AWS service integration. It's secure, stable, reliable, and easy to use, particularly by utilizing IAM roles for resource access without credentials.

I'd rate the solution nine out of ten.

I use the solution for my deployments.

Earlier, it used to be a struggle when we had different ecosystems. We did not always have purely server-based or serverless ecosystems. Every company has a combination of server-based, serverless, and container-based architecture. There are different environments, too. It would be a complex task to create a separate ecosystem for everything. If we do the same in CodePipeline, it is like a simple configuration file. We can have integrations for each service. They can be for the containers with EKS or the Lambda functions.

Compared to any other tools, AWS products provide better integrations. We have tools for all our needs. The integrations are good. The product is a fully managed service. We specify the various stages in the CI/CD process. We can do it without any external tools. Going through everything is usually an overhead for developers. It is like a configuration. We need to configure it only once.

The integration with other AWS services has helped us. Our life as a developer is easy. We need not focus on the integration manually. If we work with third-party tools, we must consider connectivity, role management, security, authentication, and authorization. In AWS CodePipeline, if we have IAM roles configured and KMS for the credentials, we need not worry about anything else. Everything can be done within the tool. The integrations are the best part. We can track everything. The connectivity and scalability are good. The architecture is also flexible enough. We can add multiple things.

There is limited room for tuning beyond the timeouts, memory, or CPU allocations. Sometimes, we need to take some screenshots of the results. The tool does not provide automated features for evidence collection. We must manually modify them to place them in S3 buckets and configure them.

I have been using the solution for two years.

I rate the tool’s stability an eight out of ten. The stability is good. I didn't see any crashes. I saw some bugs occasionally. They were minute.

The product’s scalability is good. I rate the scalability a nine out of ten.

The implementation is easy. There is a learning curve to understand the features. I struggled during the implementation the initial two or three times. Later on, it was easy.

The product is worth the money. The developers need not focus unnecessarily on the build and other processes. They can focus on problem-solving.

In addition to Maven, CodeBuild supports Gradle. Gradle builds have some issues. The tool provides more control over the underlying agent containers. Earlier, we used in-house products. I will recommend the solution if someone is using AWS. It comes with an entire ecosystem. Overall, I rate the solution an eight out of ten.

Our primary use case for AWS CodePipeline is to automate code deployment to infrastructure.

AWS CodePipeline is very useful in streamlining the deployment processes. It allows us to have multiple independent pipelines that can run in parallel if needed. The solution supports our team in managing rollbacks and version control by handling them like a new deployment. The simplicity and integration with other AWS services make it a significant addition to our workflow automation.

The management of CodeBuild and CodeDeploy on a three-point trigger basis is an important feature. The tool's capability to handle CodeBuild and CodeDeploy efficiently is significant. It's very useful for running multiple pipelines independently or in parallel.

AWS CodePipeline is quite a simple tool mostly for management and creating automation. Perhaps integration with other vendors like GitLab or GitHub could be improved.

I have been working with AWS CodePipeline for over three years.

AWS CodePipeline is highly available and highly reliable.

AWS CodePipeline is a fully managed service. It is scalable and highly available.

I have not had the need to consult support as the documentation was clear.

Positive

For Jenkins, Jenkins can handle the full flow, including build, deploy, and all pipelines.

AWS CodePipeline setup has prerequisites, but the process is fairly easy and does not have any outside dependencies.

AWS CodePipeline is free for GoDaddy, but it uses CodeCommit, CodeBuild, and CodeDeploy in the backend, which have their own charges.

I would highly suggest using AWS CodePipeline if you are also using CodeCommit, CodeBuild, and CodeDeploy. If you are using a different code repository or build system outside AWS, then AWS CodePipeline might not be the best addition.

I'd rate the solution ten out of ten.

I have used AWS CodePipeline as part of my work deploying applications, particularly an e-commerce app for a customer on AWS. I designed the process starting with building Docker images on a local machine and then pushing them to AWS ECR (Elastic Container Registry). 

After that, I configured the necessary infrastructure using Terraform and used AWS CodePipeline along with other services like CodeDeploy and CodeCommit to manage the deployment process.

AWS CodePipeline has improved our deployment efficiency, allowing us to orchestrate services such as CodeBuild, CodeDeploy, and CodeCommit seamlessly within AWS. Its native integration with AWS makes it easier to manage processes and improve deployment workflows within our organization.

AWS CodePipeline's most valuable feature is its seamless integration with other AWS services, making it easier to orchestrate deployment processes. It scales easily, and since it's native to AWS, most functionalities are integrated naturally, providing a consistent environment. The fact that alternatives like Jenkins and GitHub Actions require more effort to integrate makes CodePipeline the preferred choice.

AWS CodePipeline works effectively, and I haven't identified any specific areas needing improvement as it has been quite satisfying in terms of deployment needs. Feedback from clients has also been positive, indicating satisfaction with the service.

I have been using AWS CodePipeline for over two years now.

I have not encountered any stability issues with AWS CodePipeline since I began using it. It has been quite efficient and reliable for all my deployment tasks.

AWS CodePipeline scales easily and effectively, supporting various deployment models and workloads. It offers significant flexibility whether deploying on EC2 instances or using container services.

AWS support is excellent. Whenever I encounter issues, the support team is always available and helpful.

Positive

In the past, I have used solutions like Jenkins and GitHub Actions. However, AWS CodePipeline is more integrated within AWS, making it the preferred choice for projects hosted on AWS.

AWS CodePipeline and its associated services do not incur significant additional charges. The cost primarily comes from deploying other AWS resources like EC2 and S3 alongside the pipeline.

I have considered Jenkins and GitHub Actions as alternatives. However, the native integration of AWS CodePipeline within AWS makes it much easier to use in that environment.

I recommend AWS CodePipeline, but it depends on the type of project and needs of the customer. Based on my experiences, a rating of seven or eight out of ten reflects its capabilities. It's essential to consider the use case when recommending solutions.

The best thing about AWS CodePipeline is that we don't have to manage agents. I have worked in Jenkins for three years, but when it comes to AWS CodePipeline, you don't have to manage agents. We pay for the agents only when the build occurs. In the case of Jenkins, we have to provision a standardized node to work as an agent.

In Jenkins, we can use custom plugins from the open-source market to integrate the pipeline with any third-party application. In AWS CodePipeline, we can only use certain tools for which AWS provisions plugins. Otherwise, we have to create Lambda functions and write our own Lambda code to make those tools compatible with each other.

We recently used a DAST tool called Rapid7, for which we didn't have a plugin. We have a plugin for Rapid7 in Jenkins, but we don't have a plugin in AWS CodePipeline. We had to write a Lambda function using the Requests module to do all the authentication.

I have been using AWS CodePipeline for the last one year.

AWS CodePipeline is a stable solution.

Around 30% of our entire workforce works with AWS CodePipeline.

Sometimes, the support person is very highly qualified. They help us resolve the issue and suggest additional best practices. Other times, we encounter a support person who is not qualified enough.

Neutral

I previously worked with an open-source tool called Jenkins. I feel that AWS CodePipeline is a better solution than Jenkins. Managing the Jenkins server and nodes is a bigger headache than AWS CodePipeline.

My client is a US client, where the deployment is done in Lambda. In CodePipeline, the first stage will be source code. The source code will be pulled from GitHub using AWS CoStar. We pull code from source code using CoStar connections. Then comes the build stage, where we run the npm run command to build the front end.

We push the build artifact and index. html to S3 bucket. We host the front-end website using S3 bucket and CloudFront. Back-end applications are node-based applications that are deployed in Lambda. There, CodeBuild builds the application using npm. After that, we use the built artifact as a layer for the Lambda function.

The deployment happens in the cross-account. We have separate accounts for each environment, and we create an IAM role in all those accounts. We give STS permission for those roles. In principle, we are giving the CI account.

We have a centralized CI account that has AWS CodePipeline. In this CodePipeline, we will assume the role created in the other account. Assuming that role, we are going to deploy in that account.

The solution's initial setup is a bit hectic, but these things need to be done to maintain security. An easy deployment might result in some security lapses.

I would recommend the solution to other users. AWS CodePipeline is worth the price. AWS CodePipeline is quite easy to learn and work on.

Overall, I rate the solution an eight out of ten.

I use the solution in my company to create an end-to-end flow for the clients' pipelines and to host different applications in different languages.

The tool improves the functioning of my company and makes continuous delivery seamless. Once we set up the pipeline and you connect it to your branch, you can make any changes to different branches anytime. The tool provides an easy way of scanning the entire setup packages and the applications we work with within our company.

The solution's most valuable feature revolves around migration. The tool makes it easy to deliver the software. You don't need to sign in to any script to do anything, so you can just package everything in the tool.

AWS continuously improves the tool's UI.

One downside in AWS is that when you attempt to push a change in, it misses that part, or it could be because some variables are not set correctly. When AWS CloudFormation Stack goes for an update, it gives a failed status. Most of the time, you have to delete the stack and run your pipeline to create a new stack. You can't go back to reverse those changes with the AWS pipeline if a stack fails when you've pushed a change that didn't match up so well.

It is not true that our company doesn't have a backup solution, which will help save everything in case something goes wrong. It is not that a backup is missing but more of a configuration based on how AWS CodePipeline is set up, like how AWS CloudFormation works.

I have been using AWS CodePipeline for over two years.

It is a very stable solution as long as the developers don't add anything that will end up breaking the pipelines. If we communicate with the developers about everything we do in our company, then breakages do not really happen.

Linux is where you may have some limitations on the number of pipelines you can create through the tool. If you want to have more pipelines than the limit, you just need to add a new stack file. One stack file can accommodate a particular number of stacks, so you add a new one, then reference each of the pipelines to that stack, and then continue creating pipelines.

Everyone has their own deployment strategy, depending on their needs. But I think it's pretty popular, especially when you are doing AWS-type deployments. The tool is used by the developers, including the front-end and back-end ones, along with the DevOps engineers.

Currently, I think we are quite comfortable with whatever we have, but I think with time, we may extend its use depending on the needs of our management.

The solution's technical support responds whenever you have an issue, especially whenever you need something sorted out from their side. The engineers are always ready to help. If you have done a wrong configuration, the solution's technical support team will help you have uptime and support when you are offline for an extended time period. I rate the technical support a nine out of ten.

Positive

I use IBM QRadar.

When you are done setting it up, sometimes a dev pushes a change and then ends up missing something or adding a variable that doesn't really exist in the stack. Then it won't update, and it won't go through so that it will fail.

The initial deployment process is not very easy, but then you have to set so many things up. You need to have quite a number of files when you are setting up the CI/CD pipelines. You also need to know the approach that you want to use so that you don't end up using a strategy that will not work out very well with the expectations or with the requirements that have been set.

The solution is deployed on the cloud. You can deploy the application on VMs. For most of the production setup, it has to be a private cloud. For other test purposes, you can go to the public cloud to deal with different types of instances. If there are front-end applications that need to be accessed by the users, then it is better to go with the public cloud option as long as one does not have sensitive data.

The type of strategy we have when deploying the tool is something that will depend on the type of front-end and back-end applications. You have your branches where you are deploying from, and then you have a main area where you add all these configurations that you have added in the individual branches when you run to execute the creation of the pipelines and the deployment of the applications. For a well-prepared user, the tool is easy to deploy.

To deploy the tool, you just need to discuss it with your front-end application developer and hire a new back-end developer for the two applications that you need to deploy if you are doing both front-end and back-end there.

I don't have to maintain the tool once the pipeline is set up. As long as one observes exactly what the setup is like, which involves not adding new variables without communicating to the DevOps engineer, everything is pretty seamless.

For the maintenance part, there is a need to basically ensure that if there are any new variables that need to be added, the developers communicate to you with the actual variable or the value. With the variable names and the values, if you are doing deployments on an EC2 instance, you would need to have a look at if there are any changes in the libraries. You would need to also update your AMI. If you use AWS, then you would need to update your Amazon Machine Images (AMIs) so that when you are evaluating your pipeline, there is no clash because of the different versions of the libraries you run.

AWS charges you based on the number of pipelines you have and how active they are, and I also think that the root account user knows about all the price-related metrics. The tool offers the best value for the money one pays for it.

The configuration and setup of the tool are good areas.

For now, I think the tool has more advantages than disadvantages.

The tool is very easy to use for the setup, deployment, and continuous integration processes, so it is an effective product. There is not much need to maintain it. Initially, you need to know exactly how you want to map everything.

I rate the tool a nine out of ten.

The primary use case is for CICD pipelines. It is used to connect CodeBuild and CodeDeploy in a streamlined solution. The AWS CodePipeline only works with the code build and code deploy.

The integration with other services and notifications, along with the approval method for the production environment, helps ensure that the pipeline runs precisely when it should, avoiding unnecessary runs.

The integration with other services is valuable, just like CodeBuild and CodeDeploy. Also, the notification and approval approach in the production environment are very useful.

A bit more control over AWS CodePipeline would make it a better tool. Improved enhancements and features could make it more intuitive.

I have been working with CodePipeline for four years as an MSP and consultant.

If everything is set up properly, CodePipeline is a stable and reliable product. It depends on CodeBuild and CodeDeploy stability.

I would rate the scalability of CodePipeline as low since it is very dependent on CodeBuild and CodeDeploy. Its scalability is limited.

I have not communicated with the technical team of CodePipeline.

Neutral

I would rate the initial setup of CodePipeline as nine out of ten. It is generally easy, but there could be challenges in the first attempt.

I would rate the pricing for CodePipeline at a five out of ten. It depends on other services being used in conjunction with it. The pricing is a bundle.

I advise to go with test runs and see how everything is functioning. It's best not to integrate everything in one go. Challenges might arise while setting up initially.

I'd rate the solution seven out of ten.