Good-Bye Data Center

[Andrew Laden]

So be begins, or ends. Data Center End of life March 28th 2029 (end of sale March 2028, 2028)

So much for "No plans to discontinue data center"

We all saw the writing on the wall.

[jodywhite]

Link to announcement?

[Michal Patoprsty]

https://www.atlassian.com/blog/announcements/atlassian-ascend

[Ian]

Hi Andrew,

Hope you're doing well! I was wondering where this announcement came from? I've searched but couldn't find anything.

Best,

Ian

[Eric R Hartway]

Where did you get this information? 

[Nancy Debnam]

The announcement was emailed out.

[Doug]

Email from Atlassian or a third party? I've received nothing so far.

[Eric R Hartway]

I've received nothing as well. 

[Nancy Debnam]

info@e.atlassian.com

In an hour, Atlassian will be making an important announcement about our Data Center products. Because you are a highly valued customer, I wanted to reach out directly in advance of our public announcement. I am ready to respond to any questions or concerns you might have and discuss next steps accordingly.

You'll receive a general email shortly with the following information:

Our job has always been to unleash the full potential of teams. We released Data Center a decade ago to achieve this for enterprises, and it has met the needs of the past decade. But the things you hired us to do – supercharge teamwork, stay competitive, innovate faster – can no longer be delivered via Data Center products as we look to the future. To truly set you up for success, we need to provide so much more. On our cloud platform, you can unify your data for tailored experiences, bring AI on as a teammate, and move faster as a team with tightly integrated apps, all while staying protected against rising security threats.

That's why we've invested heavily in our data foundation as well as security, scale, and compliance so that Atlassian Cloud can continue to meet the needs of all organizations, including large, complex, and regulated enterprises. As a result, today 75% of Atlassian's highly regulated and enterprise customers are in cloud or on a path to cloud. Along with reporting faster time to value and reduced costs, these customers are benefiting from AI‑powered teamwork, moving work forward up to four times faster.

With 99% of our customers in or on a path to cloud, and demanding even more innovation to meet this moment, we are shifting our full focus to the cloud and announcing that Atlassian Data Center products will reach end of life on March 28, 2029. This focus will allow us to innovate even faster and ship even more benefits to help you stay ahead.

We know this is a significant transition. To give you time to plan and prepare, we are providing a three year transition period which is outlined in the timeline below.

You'll soon have access to a detailed timeline, transition information, and online support resources to help guide you through these changes. We'll send you a follow‑up email with a direct link as soon as these materials are available.

We recognize that every organization's journey is different, and we're committed to being a flexible, responsive partner as you plan for the future.

Our team is ready to meet with you, and are committed to collaborating on options that best fit your organization's needs.

Whenever you're ready to talk through your options or next steps, please respond directly with your availability and we can arrange a time to meet.

Best, Stuart Bennett

[Ian]

Thank you for posting the email Nancy :) 

[Volker Heubel]

Wonderful..

As an european university with sensible data we want to be in control of our data. Especially in these days… 

Two years to find an alternative and start migrating.

Bye bye

[Marco Nowak]

We are also f*cked up by Atlassian`s customer treatment. 6 months ago we were told by an Atlassian employee, that there will be no plans to end DC.

Bye, Atlassian. We might buy a tool from a different provider who treats its customers better and has better pricing policies.

[Marcelo Viana de Siqueira]

What about this?

https://community.atlassian.com/forums/Data-Center-articles/The-Future-of-Atlassian-Data-Center/ba-p/2006054

"To be clear, we have no plans to end of life or end of support for our Data Center offering."

[Gal Fatal]

That was more than 3 years ago.  Everything is changing so fast, it was clear that Atlassian will stop supporting DC one day..

[Ryan Goodwin]

Wow.

[Elias B_ Sørensen - Kantega SSO]

Official announcementa are out now: 

- https://www.atlassian.com/blog/developer/from-data-center-to-cloud-the-next-chapter-for-marketplace-apps

- https://www.atlassian.com/blog/announcements/atlassian-ascend

[Ulrich Kuhnhardt _IzymesCo_]

FYI and worth noting: Jira Align DC and Bitbucket DC are excluded! I wonder why?

[Marcelo Viana de Siqueira]

In 2022, “No plans to end of life or end of support for our Data Center offering”
In 2025, “Thank you, Data Center”

In IT, three years are enough to turn “no plans” into a farewell. Promises here come with a very short shelf life.

[Eddy G]

Extremely disappointing news. As power users of the Data Center suite, this is going to have a major impact. Some of the things we do are just not possible in Cloud.

Very sad day.

[Mirek]

"Some of the things we do are just not possible in Cloud..." - I agree.. but I would add to it at the end "yet"..

When looking back what was possible few years back the it was really bad.. but now it is I would say "acceptable" and I expect that within 2-3 years much more would be possible.. especially in a AI era right now.. 

I am impressed how Atlassian (since 2007 when first "cloud version" was introduced.. ) step by step clear all the roadblocks to be here where we are so I guess that things are possible.. 

As for decision itself .. it is nothing new..  look back what Adobe did in 2013.. it was actually pretty much the same journey.. 
https://www.launchnotes.com/blog/how-they-launched-it-adobes-cloud-transformation

Microsoft also working on O365 Cloud, Salesforce... so this is how it is.. now Atlassian joins this "family" and give us actually a 3,5 year heads up.. 

Am I happy that this is happening.. not quite.. I remember old times.. and I was hoping that we can get couple more years after that kind of sun setting DC announcement (like 5 years.. ) but on the other side there would be probably never enough time.. and from I see Atlassian is giving an option to extend this beyond deadline (so for those customer that still have problems there might be solutions in 2029 or later.. )

As for data residency and other that kind of stuff which was a big blocker.. lets wait for this Isolated / Government solutions how this would envelop..

So for all on-prem / DC enthusiast (that includes me.. ) .. Imagine simply now that Cloud is a new major version of Data Center app that you need to prepare and upgrade to (like from Jira 5 to 6 or .. from 9 to 10.. ) .. not every DC upgrade was working fine on a test system right? Not all addons where compatible.. there were bugs or unexpected problems.. Same here.. try, list all the things.. and try to find a solution.. and if you cannot reach to Atlassian or Solution Partner or Community here and maybe things would be possible when more people are looking into it. 

In upcoming months we would all learn from each other cases and I would like to see on Community here that people help each other in this transformation.. At least this is my personal mission here to learn and understand what are the the problems, or limits and try somehow to help. 

[Nancy Debnam]

@Mirek RE: "Some of the things we do are just not possible in Cloud..." - I agree.. but I would add to it at the end "yet"..

Some things will never be possible in the Cloud.  We have several closed environments and restricted environments.  No internet access.  No possibility of Cloud.

You mentioned Adobe's transformation, but they had to back pedal and come up with an alternate solution for those environments that are closed and restricted.  Their solution is clumsy and not so well thought out. 

I am just sorry to see that another great tool, that we have been invested in for more than 15 years is going to abandon us isolated environment customers.

[Eddy G]

If you run a "mostly stock" Jira Data Center instance, then sure, this is probably not that big of a deal. But that's not us. Customizing the Data Center apps is what gave them super-powers. 

Take Jira for example. We have dozens of Marketplace add-ons to provide functionality not provided by Atlassian. One of the most important is Script Runner. In Cloud, Script Runner runs in  a locked-down, REST-only world. You lose JVM/DB access, most UI surgery, arbitrary Groovy in workflow guards, and native JQL add-ins. The stuff that's still "supported" has many Cloud-specific constraints:

• UI Fragments cannot hide/remove UI items, and can only have a maximum of two fragments per location, and any content has to be "hosted" somewhere.
• Behaviours only work on supported fields only and runs via the UI Modification API. Mapping is limited to "All projects" or "All issue types", but not both.
• Picker-type scripted fields (e.g. external databases, issues, LDAP) are not available
• Custom workflow conditions and validators use "Jira Expressions" with no REST access so logic has to be much simpler than Groovy
• Custom JQL functions don’t plug into native JQL directly, you have to run them in ScriptRunner Enhanced Search, then use saved filters/gadgets that integrate with the app’s index
• Custom REST endpoints are not supported at all
• There's no arbitrary “switch user,” no direct Java API/DB/filesystem.

And that's just Script Runner (and I've only started researching its limitations).

So what some may consider "acceptable", we definitely do not.

[Mirek]

Yes, @Eddy G totally agree and that is true that many instance are heavily customized and yes mostly Script Runner was used in the past even that many of those solutions could be replaced over time by different functionality.. For example switching user that you mentioned is possible already on Cloud but was never on on-prem (and the Script Runner was a good solution that I personally also used many times.. ). To know what else a deeper understanding is required and feedback from the vendor..

Remember that now when the exact date is announced many vendors would probably focus on better planning and resolving most painful problems. Looking at ScriptRunner roadmap an seeing updates in comments like here that "we are working with Atlassian" on something and expect to be available on 2026 is promising.. So providing feedback like this above is crucial. 

If you have many addons then for sure work need to be done to contact vendors to figure out how they want to resolve those specific concerns. I have many examples that few years back couple of vendors said that they do not have Cloud version because of this or that reason (mostly some limits.. ) .. and now suddenly they have a solution! So what changed? Suddenly they figure out how to bypass things.. or something started to be possible. That is all.

So I hope that the same would be in upcoming 2-3 years.. That is my hope but yes, if you compare 1 to 1 there are definitivelly many roadblocks and workarounds.. but isn't Script Runner also a one big workaround for on-prem? Maybe on Cloud someone would create similar solution that would fill out those gaps.. That limits and lack of some functionality that we had on on-prem was an inspiration for vendors and look how many addons were created.. It does not mean that now those addons need the sam functionality on Cloud.. We have to revisit. This is what I hope we can also do resolve here on Community.

@Nancy Debnam - I totally understand and I guess you have to wait for details how Isolated Cloud would work in details but I guess that Atlassian announcing this are definitively thinking about "isolated environment customers". In today world Cloud means "someone else computer" but if this would be a heavily guarded environment, regulary updated and monitored on who view what then what else we need? I guess that many companies do not have a good security / patching policies and really good set of security tools are pretty expensive so please also look from this perspective.. Keeping things isolated does not mean that it would be secure but again 2-3 years ago it was no an option and now next year it might be.. Let's wait and see I am also really curious..

[Andrew Laden]

Interestingly enough. Buried in the announcement is this little tidbit

We are committed to developing solutions that meet the needs of all customers. We encourage you to engage with your Account Executive directly so that we can ensure our product roadmap meets your needs. We understand that some organizations face unique challenges or requirements that may delay their move to the cloud. For these situations, we’re committed to offering extended maintenance for certain Data Center customers by exception after March 28, 2029, ensuring you have the flexibility and support you need for a successful transformation. Please reach out to us to discuss migration options.

So i Guess if you have enough money, you can keep it.

[Mirek]

@Andrew Laden - Hmm.. Nobody said that it would be for additional fee.. (however Microsoft suddenly did that recently for Windows 10.. ). In Atlassian would probably be case by case solution who can go beyond 2029.. we never know what would be decided in the future.. Just to give an example did you know that Fisheye is still supported on server (not DC) version in basic maintenance mode (you were able to buy / renew license etc.) and prices did not change. 

This show that there might be exceptions and that is normal since not everyone i able to do the move exactly at specific point.. like here we already see exceptions for Bitbucket DC which is good that Atlassian is aware that it could be problematic..

So there is nothing else to do just to know what the limits are and simply use any opportunity to talk to Atlassian, vendors and provide / gather feedback. Without this who would gonna know how people use those products and what people really need? 

[Marco Nowak]

"So there is nothing else to do just to know what the limits are and simply use any opportunity to talk to Atlassian, vendors and provide / gather feedback. Without this who would gonna know how people use those products and what people really need? "

-> but there are different ways how to get these information from your customers (that pay a lot of money already) instead of announcing the end of a well used product and causing a lot of trouble and effort for them!

Not to mention the additional costs: a Cloud upgrade to the "Premium" option would currently mean a cost increase of nearly 50%! And the internal costs we would have for planning, configuration, migration, testing etc. are not included.

[Eddy G]

Absolutely on the internal development cost! 💸

For deployments like ours with over a decade of customizations, the loss of productivity (due to having to redirect capacity planning to re-architect Jira, Confluence and Bitbucket for Cloud) is going to be huge. If the functionality of the end-result was the same, it might be easier to swallow. But knowing that Cloud has limitations that will impact the functionality of how we use the Data Center apps, it makes the prospect of a forced Cloud migration extremely unappealing.

And don't forget to add in the cost of Atlassian Guard if you want audit records from Jira and Confluence to be available in the streaming events API... 💰

[Daniel Ebers]

To be honest, not so much surprised about the announcement, but...

... what I am not really getting is that just 47 days ago there was a different announcement mentioning the >strategic initiative to unlock greater scalability<

As part of the OpenSearch project—our strategic initiative to unlock greater scalability, enhance search capabilities, and eliminate indexing issues in Jira Data Center [---8<---]

Would somebody really start an strategic initiative knowing any greats results from it will only last until EoL in 2029? (Not even considered that OpenSearch is not 100% ready, yet).

So to me personally it looks like there would have been plans around keeping Data Center products for a "little bit longer" still ... but then a recent decision to go full-stop was made, all plans and intentions derailed - followed by the EoL announcement.

This is not a single occurrence - here is a second example about an improved App compatibility checker from July 2025

Why would somebody in July 2025 start building an improved checker knowing it will only be available until 2029 because EoL?

The announcements gave hope but same time everybody saw the signs.