This isn't an EOL for on-prem JIRA. It's a price increase. They used to offer two on-prem products, "Server" and "Data Center," and now they'll only offer "Data Center," which comes at a higher price. (They're also raising the price of Data Center.)
> It's On February 2, 2021 PT, we will stop selling new licenses for our server products and cease new feature development in our server product line...maintenance and support for an additional three years, ending February 2, 2024 PT.
> Even with three years to prepare for these changes, we understand that not every customer will be ready to make the switch from our server products to our cloud products. For these customers, we'll continue to offer our self-managed enterprise edition, Data Center.
Frankly, I have no idea what they'll do because so many plugins can only work with their on-prem API.
My company moved from server to Data center and didn't notice a difference from an end user perspective.
plugins that were certified for server will now complain (but should still work albiet unsupportedly) if they're not recertified for data center, and occasionally you have to buy a license for a plugin that you had for server because they made a certified version for data center, but otherwise the software looks feels and acts the same as before.
1. You'd be violating a contract by doing so, and putting yourself/company at risk of incurring significant monetary damages specified in the contract (or having your license terminated);
2. You might be committing criminal copyright violation;
3. The most obvious countermeasure is to write something to detect whether the software stops phoning home and use it to trigger an investigation.
Not for a company, just for my own testing, but I did this several years back.
It was ridiculously easy assuming you know how a java decompiler works. One JAR needed to be patched and re-compiled, which is a trivial process if you know how classpaths work.
They’re really giving the middle finger to everyone that pushed Jira to the top, huh? At the same time, it’s sad to see another self-hosted solution bite the dust in exchange for a more lucrative MRR cash cow for its makers.
Favorite self-hosted alternatives? Bonus points if they can import from Jira preserving issue numbers?
The problem is that Jira is many different thing, and have as many uses as they have customers. I doubt that there are two identical Jira installations in the world.
I also think that this is the true reason why Atlassian is doing this. The complexity of Jira has become to much, so now they will fix the problem by forcing smaller customers into a more defined workflow, while ensuring that those who require customization are also able to pay for training and consultance.
They already split (self-hosted) Jira into a few different “streams” or flavors with different features/pricing and dramatically different out-of-the-box configurations and workflows.
TPMs and Enterprise Agile consultants can’t really make heads or tails of Phabricator. It’s an engineering centric product, hostile to those who want to enforce standardized workflows and derive KPIs. Whether this is a bug or a feature depends on where power rests in your organization.
There is a catch though - we use it only as a code review platform. It always seemed to be focused on that part more than in any other. I'm not saying everything else is bad, it just a gut feeling that code review always got the maximum attention.
And another catch is that it seems they lost their momentum even in this area. I recall times when we waited for weekly status update to see if there is something we'd have as excuse for an upgrade. Now it's even not a weekly update and speed of development seems to be slow.
The Wikimedia project and some other high profile projects use it.
I did a pretty thorough evaluation once which included Phrabricator, it came out very well, especially the work flow around code review. It wasn't chosen for unrelated reasons.
I don't know why it isn't more popular than it is. It seems to be very mature, certainly better than a lot of tools that organizations pay good money for.
There's no one pushing phrabricator. No marketing means that most of the people picking a tool to use have never even heard of it, and it's not really something that a single developer can adopt and then champion internally.
Atlassian is going full Oracle. Sub-par quality, anti-customer attitude, high prices. I should buy their stocks now, this seems to be a winning strategy in software business.
JIRA Cloud is very slow (we moved from JIRA Server not that long ago). There is about a 5-10 second pause between seeing (server-rendered? I'm unsure) data for an issue and the page actually becoming interactive (ie. me being able to change a sprint or labels or components). I'm running out of my flegmatism.
It led me to start writing an Elm client for their API, and although a lot of the features that I'd need for it to become my daily driver is still missing, it's _so much faster_.
Couldn't agree more. As much as Atlassian charges it's shameful how undersized their instances seem to be.
I've resorted to "Client for Jira" from the guys who did DeskZilla. Neither is "supported" or developed anymore, but still working fine for my day-to-day (searches especially). If there's an unsupported API I haven't found it yet.
For anyone who is looking for an alternative, take a look at https://linear.app - I've used it a bunch for side projects and it's snappy and beautiful. I know the founder from our YC batch back in the day, and if anyone is going to create a really thoughtful JIRA replacement, it's these guys.
p.s. It's fascinating how long products like Jira end up existing through sheer market penetration. I don't know a single developer who doesn't passionately hate Jira, and yet it is somehow the "standard."
As jhawk28 points out this is not actually an alternative to an on premise product, so I don't get why you'd promote it in a thread that's specifically about an on premises product being EOL.
It might be an alternative to Jira Cloud, which is cool and all, but your comment is just completely misleading.
I know plenty of developers who like JIRA and not just developers: we quickly scaled adoption of JIRA to a company-wide setup, used by literally every department in our company for a variety of different processes. From this perspective Linear is not a JIRA replacement. It is a small tool for tracking software projects, but that’s it. JIRA is much more than that.
Idk why my company chose it but it's indeed bearable if you just talk to the API and make nice overviews with `jq` instead of trying to use the web interface which will trip you up eight ways from Sunday. I guess I don't hate jira after scripting against it; it's definitely more scriptable than a lot of alternatives.
The recent batch of Jira clones all seem to be missing a feature that was one of the few things that made me choose Jira in the first place: BPM style workflows with phase-specific data capture and actions / transitions.
Trello / Kanban / Agile style 'project management', gnatt / burndown charts and velocity etc. are nice for day to day views for people working at the coal-face and the never-ending war of attrition.
But a lot of the competing tools are not as elegant when it comes to modeling interlinked cross-team processes end-to-end.
You may wish to check https://pyrus.com - they are focused on no-code end-to-end processes with very flexible (and non-bloated) workflow setup, open API, and neat mobile apps (which is a must have feature for getting timely approvals from C-level guys).
> I don't know a single developer who doesn't passionately hate Jira
Well, as long as the perf aren't too bad (which seems to depends on the number of concurrent users), I don't mind Jira and neither the rest of my team, it's a tool that more or less does its job in bug and task tracking. Though we were to customize our workflow/fields to do what we want.
It doesn't allow you to track % complete, the time tracking plugin is actively awful, the VCS integration is mediocre when it works at all. The UI actively discourages reporters from filling in the description field But promotes a million other fields that are just noise...
Now i'm filled with rage whenever i use jira cloud. It's unbearably slow. Besides that, it often happens that the text area that I just clicked lost focus somehow, i start typing, AND JIRA STARTS DOING STUFF. I don't know all the shortcuts so i don't even know what I just did.
We noticed the cloud offering became consistently worse over time for the past two or three years. Just last week we switched to Clubhouse and while it’s certainly much more limited, it’s at least an order of magnitude faster than JIRA and they really seem to care about the UX.
We switched to clubhouse a few years ago. It is snappy and the UI is really nice. I particularly liked how clubhouse drives backlog curation as a largest block to smallest block. At the time, this was way ahead of JIRA. I found the team at clubhouse to be awesome too - responsive and transparent.
Ultimately we went back to JIRA because clubhouse thought differently about agile workflow than we do. At the time, clubhouse didn't support sprints other than by adding labels - but the reports didn't reflect the sprint temporal boundaries. We found it really hard to create things like a view on sprint velocity. We also didn't agree with their view on what a team is but I forget the specifics.
It wasn't that JIRA is perfect, but it was either less wrong or more right depending on your perspective.
I am a fan of the newer Portfolio/Plan functionality in JIRA which I think replicates some of the features I liked about clubhouse, in forcing a large to small PBR, but I don't use the date planning features.
This happens to me all the time too. Sometimes it’s genuine user error (I have Firefox focused instead of an Slack or $TERM) and sometimes it’s the UI being glitchy, but each time I get a sudden moment of dread as I see the page updating in seemingly random ways.
Whoever thought hotkeys without modifier keys we’re a good thing needs to be taken back to UI/UX school and retaught the basics of predictable interfaces.
>Which is as always, unclear priorities, unclear requirements.
Tell me about it. Recently had a PM ask me for a time estimate for a task called "Create reporting tool for [network] interface data from [network monitoring tool]". No requirements. No further description whatsoever. And I told the PM three business days prior that I needed more information!
I swear there is a very large segment of PM that believe they are some kind of "big idea" and "vision" dreamers and to get intimately familiar with the details of what they asking to be built is below them.
You would get tickets like "Support protocol X" or "Integrate with 3rd party Y" with absolutely no effort spent to answer the hard questions, and most of the time they don't even know X or Y beyond the marketing.
I think in part it is because "PM" is such a nebulous role, there isn't a clear career track for someone to become a PM so you end up with a widely disparate set of skills in that role.
The best PM I worked with did comp-sci, decided he hated programming and went into project management, he got it.
The worst was promoted out of customer support and didn't have a fucking clue.
I swear there is a very large segment of PM that believe they are some kind of "big idea" and "vision" dreamers and to get intimately familiar with the details of what they asking to be built is below them.
Those same types also believe that if only they can add enough custom fields and statuses and mandatory workflows the project will become self-managing and they can go and do whatever it is they think they do.
Yes and now you can’t own the hammer, you have to call a cloud provider to bring you one each time you need to use it, assuming your hammer subscription is paid up.
What do you prefer? I've tried a bunch of products over the years and they all have some sort of really annoying flaw or shortcoming. Jira at least has easy to reference issue numbers and can be configured in pretty much any way.
Haven't used GitLab but I found GitHub's issues extremely limited and the UI turned away any non-engineers. So GitHub becomes the "engineering internal issues tracker" while everyone else uses something else. So now engineers need to keep track of two different systems and keep them in sync.
Cloud offerings that hook into your existing IAM are really good these days. In fact, even the IAM can be off-prem (Okta). It relieves the maintenance and upgrade burden significantly.
Yet, open source + self-hostable services have found a niche. Many organizations can't use off-prem for anything remotely sensitive, so they turn to software like GitLab and Mattermost. Interestingly, I have noticed that these self-hosted services broaden their scope to incorporate functionality that might be covered by another service, to reduce said on-prem maintenance burden.
This sucks hardcore. Essentially, Atlassian holds customers with confidentiality requirements for their data hostage - no support after 2024 for on-prem at all.
The key thing to remember is that Atlassian is an Australian company, they are bound to Australian laws, their cloud is not in the legal area of the European Union. For anyone who has sensitive information to store, all Atlassian projects just have died, because the vendor cannot be trusted to not abuse their position any more.
Hopefully alternatives will be rising, it's badly needed (especially given the ... shoddy state of the code/architecture behind Atlassian projects).
... This is going to mean changes in current $DAYJOB, as we are essentially locked to hosting everything in Germany.
Hosting in Australia for business unit that isn't there is right out. We will have more contact with AWS China than them at this point, because we are apparently going to have an unit there...
Among the contenders, it seems this is the trend. Microsoft’s DevOps Server is barely getting any updates compared to their cloud offering and seems deliberately held back in order to make customers switch.
Maintaining the on-prem use case for huge web+db apps is becoming too expensive as fewer and fewer customers use it.
If there is a good cloud offering then some customers will want to use it. It’s convenient, after all. Even if it’s just say 50% preference, it means you are maintaining on-prem for half as many.
Whether they drive people off on-prem by raising costs or people leave because costs are raised isn’t really an interesting distinction, it’s the two sides of the same coin.
"Serious" as in "I don't want to store my files in some random cloud".
Jira is, among other things, a document storage solution. Security aside, at the minimum most people would like to be able to back up and access their critical files.
I don’t mind storing my source in the cloud (although if you asked my employer ten years ago it would have been a different story. That ship sailed. Even the biggest enterprises seem pretty happy to use cloud solutions for most things. I agree backup (and off ramps) are critical in any cloud offering. Can’t be in lock-in and also can’t trust their backup only of course.
Jira to me is just an issue manager though. I know there are plugins these days for document management inside Jira itself but I haven’t tried them. Don’t people normally have separate document solutions like Confluence or SharePoint for docs (aside from document that are attached to issues of course)?
People who buy into Jira heavily use it primarily as a communications platform, to structure and make searchable the stuff that would otherwise be lost in the swamp of email and Slack and face-to-face messages.
> aside from document that are attached to issues of course
Aye, there's the rub. The documents in Confluence are often less important than the documents in Jira. You commit the paper trail to Jira - the primary info you get from your customers and bosses. Confluence is just the docs from the legacy crap you forgot to update last year.
Ok... So I'm trying to see what options we have now. We have Jira, Confluence and Fisheye installed on server, plus a free GitLab for code repository. We use FishEye for code revisions (we have multiple repositories and usually changes came across 2 or 3 repos at same time).
How integrates Jira Cloud with Fisheye ? I don't see a cloud option for FishEye, so it keeps being on a local server...
Strange decision. Our organization has several both server and data center deployments, and periodically there is talk about “should we switch to azure devops” where the decision always comes down to stick with JIRA due to organizational momentum. I am certain that this decision from Atlasssian will lead us to fully abandon JIRA Deployments within a year.
While we’ll just use the data center version, we have customers who will be forced to abandon Jira due to legal or contractual requirements, because they only require maybe 5 or 10 licenses, but can’t use the cloud version.
I have to assume that Atlassian have some internal number regarding how many customers they expect to lose.
Same thing: EoL for new on-prem Server licenses in Feb 2021, then end of support for on-prem Server in 2024. Only Cloud or Data Center going forward—-they said that’s their (horrendous) strategy for all their stuff.
Time to find a replacement for that one, too. Ugh.
Confluence going on cloud is much worse than Jira, as working with Confluence depends on multiple frequent successive edits whereas Jira is usually just "visit one page, do one task". On-prem Confluence is as good as working in Microsoft Word -- a desktop application experience -- while Cloud is like working in Windows 3.1, with a little hourglass cursor to wait for between every action you take. I'm glad I never migrated my notes to Confluence.
Also, they're going to lose secure government contracts with this, because Data Center is way above the size of a project team.
I dislike Jira. My team had used Trello for four years. Now GitLab is doing all the things and it's super easy API made it very easy to move stuff and integrate with my CIC and OPS tools.
Switched to YouTrack from Jira and loved it...until there was a reorg and they forced on Github Issues which are just...primitive and nobody but me knows how to use it. FML
I've bought a license for it yesterday because of this. I don't have any use for it at the moment, but there have been many cases where a tool like Jira would've been useful to automate some processes I used to do manually (and those don't warrant the time investment of building a custom webapp), so I'll keep it just in case such a need arises in the future.
I'm not the one maintaining Jira at our shop, but we've got a number of integrations. FishEye for one but also custom stuff like our customer database (so we can assign customers to issues), and in our build process (issues gets updated with built version automatically).
Not sure if those integrations would translate well to the cloud version.
There's also something about owning your own data...
Besides, I found Jira ok enough speed-wise, but judging from all the comments on HN each time Jira is mentioned, the cloud version seems not fun at all.
Business requirements (usually tracing to contracts or law) to have real stewardship over your data.
Not every product domain can operate where they only rent access to the authoritative copy of their records. Or don't have control to acceptance test / rollback the specific versions of the software involved in displaying/modifying those records.
In the past I've worked on safety critical software. As a first order approximation: either the local IT department's disaster recovery plan can ensure continuity of all the quality systems' digital records, artifacts and tools, or you don't actually have a quality system.
Speed (the cloud version is extremely slow and I can't do anything about it, on-prem gives me the option of throwing more hardware at it) and privacy/data protection.
Microsoft may have services in the cloud and may be getting rid of macOS office, but they still have on premise Windows and Office.
Similarly Oracle will continue to sale on-prem DB.
Real products can make real money. If an ISP goes down, company could internally function with on-prem.
Don’t take that away for Jira. I love Confluence, but it or Sharepoint could be replaced by a file server if needed.
Jira’s not like that. There are too many things it does where it relies on being an extendable and searchable database with a custom workflow. There’s not a hack for that, it’s an essential product.
If you can’t afford to manage it being used that way then hire someone to do it.
