Webinar Q&A: Supercharge development with Bitbucket & Atlassian Cloud

[#1]

Hi everyone, thank you to all who signed up for the webinar on Dec 5th/6th: Supercharge development with Bitbucket & Atlassian Cloud - here is the recorded version if you missed it.

We had many questions that we didn’t get to or didn’t know the answers to during the webinar and are publishing this blog with Q&A responses from our broader team.

Few things to note:

• We had several questions around upcoming features, AI in particular, that we don’t have details on yet. We’ve grouped questions around this.

• We also had a lot of CI/CD questions. We’ve answered most here.

There were a handful of questions that we weren’t able to answer objectively. Please ask the discussion type questions on the community and our team will engage on those.

Thanks again for joining the webinar. 

We've grouped questions into 3 buckets - AI & roadmap, CI/CD, Other.

Questions around AI and roadmap:

What Languages will be supported by AutoReview?

Can we instruct the AI reviewer to avoid giving us wrong errors?

Will Autoreview support different/custom coding standards?

Will it be possible to talk to Autoreview Agent so it can provide another answer/code samples based on the user feedback? Will the autoreview agent have context of whole repository and codebase or only changes?

Will it be possible to integrate self-chosen LLMs with Rovo in Bitbucket?

Will the code suggestions feature log auditable information so that we can identify all instances of generative AI-created code?

I'm really excited about the auto suggestions feature in the Pull Request. When is it planned to be released to the users?

Will AutoDocumentation also work for Documentation in Format of Markdown inside the repository?

What about pricing for these the AI improvements?

We are in early stages of planning for Auto Review and Auto Documentation. However, we will take these questions as input into our planning process.

We don’t have specific launch dates or pricing yet. We’ll update our public roadmap when we do. For our existing AI features, they are available in our Premium plan.

Our IT security is quite strict on AI usage, are their resources we can send to them to better understand the data sharing and usage from our IP?

This page should provide you with the information you need. If you have more specific questions, you can contact our support team or post on the Atlassian Intelligence & Rovo community.

My organisation has a policy restricting our use of AI tools. Can AI-specific features be disabled?

Yes, you can deactivate AI via workspace settings.

Bitbucket Cloud is lacking 2 important features: 1. Notifying the pull request owner if there is a conflict in auto merging to higher branches. 2. Force a mandatory number of pull request reviewers. Will any of these feature be available in the near future?

1. We have a public feature request for this: Please vote and comment with more details about your use case!

2. This is available. You can specify the minimum number of approvals for a pull request by configuring branch restrictions. Learn more.
   Restrictions can be set at both the project and repository level. Premium customers can make branch restrictions mandatory.

Does better execution times include extending self-hosted runner time limit beyond 2 hours? There's a feature request ticket for this as well: 

Yes - step execution time has been extended to 12 hrs for all builds. Thank you for linking to the BCLOUD. We’ve commented and closed it as shipped.

When is support for ARM builds coming?

Support for ARM builds is scheduled for Q1-Q2 next year. Here is our public roadmap link.

I was thinking you mentioned "Draft pull requests" would be a feature added in the coming year, but I don't see it in the roadmap.

We will be adding draft pull requests to our public roadmap in the near future. In the meantime, you can follow this public feature request for the latest updates on status.

When will Atlassian Guard and Bitbucket Cloud integration for group management be available?

We have this integration available today for new workspaces and plan to release this for existing workspaces by the end of next year. Here is the link to the feature on the public roadmap.

If you are considering creating a new workspace today, this integration is already available. You will be able to manage users in Guard.

Is there a plan to allow SSH into build container?

This is something we are considering as an enhancement for Pipelines in the future, but do not have anything concrete to share right now.

It would be great to understand the use-case you. Could you please share any feedback you have via the Bitbucket Pipelines community space.

Is there a way to "approve with suggestion" ?

We are in the process for planning out a Code Suggestions feature and will add it to our public roadmap in the near future. In the meantime, you can follow this public feature request for the latest updates on status.

When is Rovo slated to launch?

Atlassian Rovo is already live. In the coming year, we are working to integrate Rovo features directly into Bitbucket e.g. Autoreview is powered by Rovo. We are also planning to add Rovo chat inside Bitbucket Cloud.

Is there going to be a way to generate a pipeline yaml (ai-generated) file through new bitbucket AI features?

This is something we’re looking into in order to give customers the ability to use natural language to enhance their Pipelines workflows.

We don’t have specific dates we can share on this at the moment.

CI/CD questions:

When it comes to automating project, repo creation and doing other configurations such as permissions and branch restriction, does bitbucket have any particular recommendation on how to go about it. And does atlassian have or plan a terraform provider?

We don’t currently provide a Terraform provider for configuring Bitbucket Cloud directly - however that’s a really interesting idea and something the product team will discuss.

If you’re looking to have consistent branch permissions across repositories, one piece of advice is to set these permissions at the project level. More info

We’ve also just shipped the ability to implement schemes of Custom Merge Checks at the workspace level, and have them cascade down to every repository in the workspace.

Project-level branch restrictions are automatically inherited by repositories in that project and are applied to any new repositories created in that project. You can definite additional restrictions at the repository level if desired.

As for automation, we have API endpoints for setting repository-level branch restrictions. More info

We have a public feature request for API automation of branch restrictions at the project level.

When we run jobs in Jenkins, we rely on agent environments, which could be complicated, like compilers, test tools, requirement for various versions of windows, flavors of Linux. Is there a way to do it with Pipelines?

• For linux, this can all be achieved using Docker. Customers can either use off-the-shelf Docker images from Dockerhub, or they can even create their own custom Docker images to give themselves complete control over the environment.

• For other OS’s like Windows, these are currently supported via our self-hosted runner, which gives customers total control of the environment.

• In the future we are planning to support Windows natively in the cloud, which should mirror our linux setup in terms of how it operates.

We are using NAnt task in Bamboo, does pipelines support NAnt?

• Yes, but with some nuances, depending on how your builds are setup.

• If you require your builds to run in a native Windows environment, then you will need to use a self-hosted Windows runner.

• If you’re running a build that is able to be executed in a Linux environment, that can be done in cloud.

• We are planning to support builds running in a windows environment inside our cloud platform in the near future.

I think 'auto-merge failure notifications' is referring to cascading merges, which aren't available in Bitbucket Cloud?

We have a public feature request for automatic (cascading) merges:

Please vote and comment with more details about your use case!

We manually publish our websites, and copy the files to a hosted server. Do we need to move to Azure or AWS before we can implement pipelines to auto deploy?

Pipelines can be easily scripted to do copy/ftp based deployments to any internet exposed host.

You can either do this manually by writing the relevant scripts in a step within your Pipeline.

However we would recommend using one of our pre-built Pipes for this task instead as that is much simpler.

SCP Deploy Pipe guide

SFTP Deploy Pipe

How do you support regulated environments (e.g., healthcare) in providing guarantees of replicable (and validated) build environments if it is not possible to know or control the adopted Bitbucket Cloud version?

Every build run inside Bitbucket Pipelines is executed within an entirely ephemeral microVM, the “environment” of which can be entirely specified by the customer.

The microVM “environment” is defined as a Docker image, either that we provide, that the customer specifies from Docker-hub, or that the customer builds themselves and provides via their own private Docker image registry.

This gives the customer 100% total control over the execution environment of their build, guaranteeing reproducibility.

Customers define the environment they want to use inside their config-as-code .yaml file stored and version-controlled in the repository. Customers can also use tools like Dynamic Pipelines to ensure very repo in the entire organisation is using an image from a pre-approved list so that the entire company is guaranteed to be compliant.

Are there also tools/demos for CircleCI to Pipeline migration?

We’re planning on releasing a CircleCI migration guide in the next few weeks. Keep an eye on our migration guide page for updates.

In terms of our automated migration tooling, initially we’re focussing on Jenkins and Bamboo, but do plan to expand this out to other 3rd party CI tools in the future.

Will there be outage if you modify your bitbucket pipelines .yml file and deploy to live server?

Not 100% sure what is meant by “deploy to live server”. Pipelines is an entirely managed SaaS offering, there are no servers to manage, so there’s definitely no outage if you change your .yaml file.

Pipelines will execute whatever .yaml file is present on the commit that the build is running against, this means that if you change the .yaml file in a branch and then run a build against that branch, it will use the .yaml file that was updated by the branch without any downtime.

In case I’ve misunderstood the question - if you’re asking about doing something like a red/green deployment with Pipelines, that’s absolutely possible.

You would achieve this by having your deployment execute to your “red” environment in one step of your Pipeline, then in the subsequent step (which will only run if the previous one is successful), you would switch the traffic over.

You would ideally do some smoke tests against your “red” environment to confirm it’s functioning as expected and then execute a task to switch traffic from your “green” environment over to the newly updated “red” environment - avoiding any downtime.

If you have more specific questions around this, please ask in the pipelines community.

For self hosted agents, can we have the Stages to be persistent using a single checkout?

This depends on whether you’re using the Docker runner or the Shell runner.

The Shell runner executes directly on the machine that it’s installed upon, so the disk that it executes against is entirely persisted. Note that this will have significant downsides due to the increased risk of build inconsistencies due to environment pollution.

If you’re using the Docker runner (this is the default whenever we refer to “self hosted runners”) each step in Pipelines is a totally ephemeral environment, meaning by default we assume the repo needs to be cloned in as a first step for each step.

You can disable this however following the instructions here. That will stop the git clone from occurring on every step, but it won’t give you persistent storage between steps.

To achieve something approximating this, you would need to use something like an artifact to bundle up the relevant directories from the build environment at the end of each step and pass them to subsequent steps.

Note that, as a general rule, we don’t encourage this pattern. Artifacts are intended for things like transferring a build artifact from a “build” step to a “deployment” step - not for moving the entire repository between steps.

Suggestion: Is it possible to have a dashboard view for the pipelines? Like a dashboard to view running pipelines, failed ones, success rate etc

This is already available at a repository level, but not at an organization-wide level.

We are looking into options to provide higher-level views of Pipelines operations, but don’t have anything we can share right now.

Why did the check-pipeline page, where we could copy our script for syntax check, disappear? Or if it has moved, where it is ?

The syntax validator is available here.

There are also VSCode and IntelliJ plugins that support real-time linting etc for Pipelines.yaml files.

Is there a way to automatically send a message to a Microsoft Teams chat when a bitbucket pipeline finishes, including also the changelog connected to the release?

Yes, you can use our MS Teams pipe in your pipeline to send a message to your chat at any stage of the pipeline. 

Do you have/plan to have tools for automating manual procedures that we need to do after deployments directly in the deployed platform?

Without knowing specifically what is being referred to here, it’s difficult to comment.

You can already have Pipelines that continue on to do post-deployment processes such as testing, environment cleanup, configuration of the deployed services, etc - you would just do this as part of the overall Pipeline in a step after the deployment. Here is more info.

Are there any plans to improve pipeline integration with iOS projects. I mean remove the need for a MacOs runner.

iOS projects require MacOS build environments in order to execute. Due to licensing restrictions around MacOS and MacOS instances, it is very difficult to offer MacOS build environments at a price-point that would be appealing to customers.

Due to this, we have no immediate plans to offer hosted MacOS build environments. However we are working on options to make the self-hosted Mac runner more scaleable.

Do any of the new features you mentioned and today including the AI Roadmap extend to customers that will continue to use Jenkins with bitbucket. We will remain on Jenkins indefinitely and may need to consider another platform for repos unless this integration is sufficient. Thanks.

You can use Jenkins with Bitbucket Cloud or Data Center. However, our AI features and our roadmap only extend to features within Bitbucket Cloud.

Jenkins may have AI features of their own that you can explore.

For PR auto reviews. Can you talk more about this? We are hoping AI can help with finding potential security issues ie not escaping SQL and potentially helping optimizing code for example performance

Autoreview is still in development. It’s intended use is to help identify syntax issues and suggest opportunities to optimize your code.

For security scanning purposes, you can use our security pipes in your pipeline. See this blog for details. 

When you use these pipes in your pipeline, you can access the code scan reports within your pull request view so you can review before you merge.

About help inside Pipelines steps, if there are errors, how deep do you think AI can go inside the customizations?

We are still planning AI functionality for CI/CD and do not have details as of now. We will share more information when we have it.

Will the maximum allowed start wait time for parallel steps be extended?

Maximum step execution time was recently extended to 12hrs, from the previous maximum of 4hrs.

Other:

What information or metrics told you that developer productivity increased by switching to Bitbucket Cloud?

We’ve been measuring NPS and PR cycle time internally for the last few years. Teams that have switched to Bitbucket Cloud have higher NPS scores for lower PR cycle times.

How can we move from AWS code commit to Bitbucket?

This blog post outlines a process to migrate from AWS CodeCommit to Bitbucket Cloud.

I want to learn on how to start using Bitbucket.. is there a foundation video?

Yes, we have a demo that walks through our core features and how Bitbucket Cloud works with the rest of Atlassian tools.

You can also access our help documentation or ask specific questions here in the Bitbucket Community.

Where do you post blogs about Bitbucket and updates in relation to roadmap?

We publish new features and roadmap updates on our blog and we also have a our public roadmap which is a snapshot of what we’ve recently shipped and what’s coming with timing estimates 

In terms of moving from a Gitlab to Bitbucket Cloud, Jenkins, etc - does Atlassian have support with this migration? Or would we need to do so via a partner?

We don’t have a tool to migrate from GitLab to Bitbucket Cloud. If you need migration help, you would need to go via a partner. However, our support team can help you answer any Bitbucket related questions as you go through with your migration.

If you’re using Jenkins, you can use our CLI to help you convert your Jenkins pipelines into Bitbucket Pipelines yml. Learn more

I would love to look at Atlassian Analytics, however our development team is only 11 users so we can't use Enterprise :( Are there any plans on making this available to lower subscription tiers? Thank you :)

Hi! There was a slide about "Atlassian Analytics: with Teamwork Graph". Is this feature already available for Premium customers? Or when is it planned to be launched?

Atlassian Analytics is only available for Cloud Enterprise customers of Jira, Jira Service Management, or Confluence.

If you have any Bitbucket workspace connected to the Jira Enterprise edition, then you’ll be able to see connected metrics across products as shown in the teamwork graph.

For teams not on Jira Enterprise, we are working on building  enhanced custom reporting capabilities next year based on the Teamwork Graph. We’ll share more on our blog when we have more details. This keynote from our founder (plays at the 4 min mark) has some more context on the Teamwork Graph and how the data enables AI as well.

When will it become possible to include Bitbucket Cloud and Trello Cloud invoicing to the main Atlassian billing?

We have plans to onboard onto Atlassian’s centralized billing platform. Stay tuned for a roadmap update announcing the timeline for this work.

Could you please elaborate on the capability of the Score Card if possible?

Scorecards are part of Compass, our developer platform. You can learn more about Compass scorecards and sign up for a trial here https://www.atlassian.com/software/compass

Where could we find the resource links after the webinar?

Here are the links we shared during the webinar.

• Bitbucket Cloud demo

• Code security pipes

• Dynamic CI/CD pipelines

• Atlassian Rovo

• Start a cloud migration trial

• Download the cloud migration assistant

• CI/CD migration from Jenkins

• Open a ticket for migration support

• Survey link

There were a few questions that we couldn’t answer. If you had a question that you don’t see on the list, please start a new thread and someone on our support team will engage.

If you want to go deeper into a specific question or have other questions, you can start a thread here in the Bitbucket Community.

For Rovo or general AI specific questions, you can reach out in our AI community.

[Saxea _Flowie_]

For those looking for merge cascade and draft pull request on Bitbucket Cloud, we provide Flowie, an addon that supports both and more!