Introduction
At AWS re:Invent 2024 in Las Vegas, AWS and Gitlab have announced a collaboration, especially with the focus on getting Amazon Q developer available within Gitlab.
In this blog I’m going to look closer at the announcement and about what I think this means for existing AWS DevTools like CodePipeline, CodeBuild and especially CodeCatalyst.
TL;DR – What has been announced
On december 3rd, 2024, Matt Garman, CEO of AWS, announced a collaboration between AWS and Gitlab.
This announcements makes AWS Q Developer with different features available with Gitlab and strengthens the relationship between these two partners.
Why Gitlab?
One of the first questions that came into my mind was “Why Gitlab?” and why a collaboration? With Github being one of the most frequently used DevTools, why is AWS interested in collaborating with Gitlab?
My personal opinion on this is that – with Gitlab being prelimary focussed on building strong DevTools that cover all parts of the product lifecycle, the relationship might be a win-win for both of the partners. Github is already too strong, too big and they most probably do not “need” AWS to collaborate as their platform is already atracting organizations and individuals more than enough. Gitlab on the other hand has been challenged by a couple of things in the past months (and years) and will benefit from some additional traction comming towards them.
In fact, I’ve heard rumous about Gitlab looking for investors – and we don’t know what happens “behind” the official announcements…
Gitlab is as an organization also flexible and agile enough to be able to react on changing demands of AWS customers that might come with this collaboration. With the focus on Open Source, Gitlab can also have AWS teams supporting changes in their code base.
What does this mean for AWS DevTools?
Well, “It depends” 🙁
What are AWS DevTools?
Let’s first define what we see as “AWS DevTools”. This itself is a difficult thing as the view on it it different. I personally count CodeCommit (deprecated), Cloud9 (deprecated), CodePipeline, CodeBuild, CodeArtifact, CodeCatalyst and ECR into the “DevTools” category, but if you look at things a little bit broader, also Amplify, AWS CDK, SAM could be seen as part of the “DevTools”. The only one of these that offers integrated, end to end tools for the product lifecycle is CodeCatalyst. As you most probably know, this is has been my favorite service for a few years.
DevTools at re:Invent 2024
If you look at the re:Invent session catalog, however, there seems to be a pattern of “services that get or do not get love”. Unfortunately, I have not been able to find a lot of sessions on the AWS DevTools in the catalog. Especially I have only found 3 sessions with a mention of AWS CodeCatalyst – which is a pitty, as most of the features announced for the Gitlab integration where already available in CodeCatalyst in 2023. This was totally different at re:Invent 2022 and 2023.
So, what does this mean?
CodePipeline, CodeBuild and CodeArtifact are essential building blocks and most probably also used insight AWS intensively – but they do not “compete” with the Gitlab integration, as CodeCommit & Cloud9 have been deprecated.
Because of this, I do not expect that this new collaboration will have a bigger impact on the development of these services.
Now, for CodeCatalyst, I am not sure.
A lot of questions and as I already wrote in a previous article, CodeCatalyst did not have any major announcements in the 2nd half of 2024. This also means that it unclear if the new functionalities, that are now available in Gitlab, have also launched in CodeCatalyst.
As I explained with someone from the CodeCatalyst team in this video, the way that the /dev
feature is implemented in CodeCatalyst has a backend that runs Bedrock underneath. I assume that the same or similar backend services power the Gitlab and the CodeCatalyst implementation, at least that’s what I personally would do. I will need to test and verify if that is correct.
Still, without major updates and announcements, it’s unlikely that there is active development into CodeCatalyst currently, as the expertise to build DevTools at AWS has always been….let’s call it… “small sized”. So, the next weeks and months are going to decide about the path that CodeCatalyst will take.
Are you an active CodeCatalyst user? Please reach out to me and share your experiences with me!
Why I am disappointed of the announcement?
Maybe I am judging on this collaboration too early, but hey – an infrastructure and “building blocks” provider like AWS now “integrating their services in a 3rd party provider”? This sounds – a tiny bit – odd for me and I am not sure what to expect next. AWS is entering the space of building software & tools for developers, but without being able to control everything end to end – like they would be able to with CodeCatalyst.
If you are a subscriber to my YouTube channel you might remember that I, after the deprecation announcement of CodeCommit and Cloud9, tried to deploy “integrated devtools services” to see what could be used as an alternative to CodeCommit. I managed to get things deployed for two other tools, but for Gitlab I never published the video – because, after spending hours (and days) on it, I gave up – I didn’t get it to run properly on ECS and I did not want to pursue the EC2 path as suggested by the Gitlab documentation.
What I am trying to point out is that I would have loved to get a “managed service” to stand up Gitlab in my own AWS account, supported, mantained and managed by AWS. This would have made a huge difference in terms of how I look at the collaboration between Gitlab and AWS. It would have looked like a complete partnership, enabling AWS customers to use Gitlab as an integrated DevTool.
Also, it would have given AWS the power to control the infrastructure and network connectivity for the Amazon Q developer features that are now available through Gitlab.
What’s next and what stretch goals do I see?
If the integration between AWS and Gitlab is meant to “fly” and create additional traction to AWS Q Developer, the AWS team has some homework to do. I already mentioned the “managed service” dream that I have, but I also would encourage additional integration options with AWS from Gitlab. What about bringing integrations between Gitlab and AWS, with certain aspects of the AWS console or other DevTools?
What about a possibility to convert Gitlab pipelines to CodePipeline V2 on demand?
What about accessing AWS services and the verification of “Drift” against deployed AWS resources?
There is way more things that could come out of a closer collaboration between AWS and Gitlab!
And now, what is a “AWS DevTools Hero” in 2025?
If I look at my role as a DevTools Hero, I tend to get a little bit nervous as I look at the recent developments. What is a “Devtools Heroes” at the end of 2024 and the beginning of 2025? Should I become a “Q developer expert” and give guidance on the best prompts for Q developer ever? Or should I rather focus on CodePipeline or AWS Amplify?
What do you think should the role of an AWS DevTools Hero be in 2025?
Please let me know in the comments!
Some tasks to do after re:Invent 🙂
Now, reflecting after re:Invent 2024, I believe that there is a bunch of things that I should look at. Not promising if I have enough time to do all of it – but I think I should
- Look at the current functionalities in Gitlab and review how they work
- Discuss with the AWS teams to find better options on integration
- Set up Gitlab 🙂 and enable Q Developer in my own account
- Plan a migration strategy for all of my projects “off” CodeCatalyst?
Feedback?
Do you have feedback or thoughts around my thought process? Please let me know in the comments or reach out to me on LinkedIn.
Views: 98