Today we’re announcing that Forge will continue to remain free until December 31, 2024, extending the free period for Forge an additional year.
Forge developers can continue to freely innovate on the platform and focus on solving problems for customers as you ramp up with Forge, without the overhead of platform costs.
We’ve published a post on the developer blog detailing this update and previewing what we’re evaluating for future pricing. Please post any questions you have here in the community.
Should be free forever. My entire infra costs are ~$5/mth to serve 364 Connect apps. I pay $0 for hosting/bandwidth and $0.50 per million backend requests (Cloudflare Workers).
Forge must be equivalent or cheaper than what’s on the market, otherwise why even charge developers?
This strategy sounds like a bait-and-switch program, entrepreneurs need to know expected costs in advance to know what to optimize in terms of bandwidth, storage, etc.
@srajan is this going to be paid EXTRA, apart from the cut Atlassian already takes from vendors based on marketplace revenue sharing? Or will this be incorporated into the revenue sharing, so cost for vendors will remain similar after this change?
Hi all - I see a few common themes in the comments and wanted to wrap my answer into a single reply. I hope this helps clarify.
In regard to @RadoslawCichockiDevi’s question - we are still evaluating future revenue share rates, and will be taking into account partners’ Marketplace costs holistically. That means that as costs rise in one area, like Forge usage, they may decline in other areas, like Marketplace revenue share.
I also wanted to highlight that future Forge pricing will follow a freemium model, with a generous free usage tier. Unlike AWS, Forge pricing may not be fully usage based, and we are in the process of evaluating pricing models. While it is too early to define what the future pricing will be, directionally, we plan to price Forge in such a way to make it an attractive option for partners and conducive to business growth.
We understand that partners need to know expected costs and we will provide at least 6 months notice before pricing changes take effect (more if possible). However, you can expect regular updates on our thinking in this area as details become more clear, and we will be working closely and iteratively with partners along the way.
Let’s be perfectly clear here and not beat around the bush: Forge is not an “option” for vendors. Although there might not be an end date yet, it is well known within the Marketplace Partner community that Atlassian has every intention to sunset Connect.
There is no need for a pricing model, as there is nothing to compete with. If Atlassian cannot make Forge work within the context of the revenue share model, you will seriously need to re-consider your strategy.
This particular sounds really scary and messy to me.
I get that it is tempting to say “we’ll use revenue share model to have partners collectively pay for basic infrastructure and X free tier and add a usage-based pricing model to prevent abuse” but that is just going to make it very fussy.
Also, although I fully understand that Atlassian can’t compete with other hosting solutions like Cloudflare in terms of pricing because the number of customers for Forge infrastructure is pretty limited as it is only geared towards Atlassian developers and not… like… the entire world, I have to be completely honest: that’s on you. You choose to remove alternative options from vendors by promoting (and highly likely eventually forcing) the use of Forge.
Just for future reference Remie, you can omit this part of your answer. If you start talking about the weather in an answer we’ll know your account has been hacked
Against better judgment I keep adding this. Not necessarily to excuse myself for being direct, but in the hope that some day Atlassian will skip the formalities and niceties. No need to pretent that the decisions of the company are in the best interest of others when in fact they are always solely in the interest of Atlassian itself. At best, as Marketplace Partners, we can hope that the interests of Atlassian also align with our own.
Not that I blame them BTW, that’s just how capitalism works, but there is no need to sugar-coat it either.
Anyway, off to another day of rain here in The Netherlands. Must be great to live in a country with great weather…
Greetings. I am new to this group. I’ve developed using both Forge and Connect. I am interested in the new costing for Forge - however I am even more interested in your statement “well known within the Marketplace Partner community that Atlassian has every intention to sunset Connect”. I have a TON riding on Connect - many important Apps. Is Atlassian really planning on “sunsetting” Connect (I imagine that means discontinuing it.). First - is this true? Second - how would that work for existing Apps? I would like to have someone from Atlassian comment about this.
I would not hold my breath waiting for an answer from Atlassian. This is a highly sensitive topic. But there are some public pointers that showcase the future of Connect.
The first is the announcement of Unifying Atlassian Connect and Forge. That announcement caused such a scare that Atlassian had to write an update. However, the update is still clear on the endgame:
We heard some developers express concern that Connect will be removed from service in 2022. We want to be clear: this is not the case, and we apologize for the ambiguity. The end of 2022 is the date when we expect that all apps will be able to start the process of going from State 1 to State 2 (and many will be able to start the process sooner). It is not a deadline for making the updates necessary to move to State 2, nor is it the date when any Connect services will be switched off.
In the spirit of OCNB, we do want to be transparent that remote iframes and JWT auth will be deprecated in Connect at a future date. We don’t yet know when the official deprecation date will be for these Connect features, but we are committed to providing equivalent functionality in Forge and plenty of time to make updates before any Connect functionality is removed. We know replacing these features is likely to be a significant undertaking for Connect app publishers, and because of this, we decided to provide this information early, even though the pathway to reaching State 2 is still being built.
If you use JWT and remote iframes in your Connect app, we ask you to collaborate with us, keep providing your honest feedback, and share the details of your implementation. Our priority is helping Partners get to State 2 with as little disruption as possible, and to do this, we need to understand your current architecture. This will inform both what we build for you and when.
Finally, new features and products will only support Forge. This is true for extensibility of Automation for Jira and for Compass (see https://go.atlassian.com/compass-developer)
Hi @bioinformatics - I’m Head of Ecosystem Platform, leading the team that builds both Connect and Forge.
The bottom line is that we’re committed to supporting Connect builders, and we absolutely will not discontinue any Connect services until we have an equivalent or better option for partners to move to - and a viable path to get there.
We know that there are many partners like yourself who rely on Connect, and maintaining stability is important. Over time, Connect will evolve so that Connect apps can adopt features of Forge, like the Forge UI, and eventually even Forge storage and compute. This is still far out in the future, however.
In full transparency, we view Forge as the future of our developer platform, and we are delivering new features in Forge that may not have a parallel in Connect. We spoke about how we make those determinations during the last Forge webinar. Although some new features may be Forge-only, we are building toward interoperability between Forge and Connect so that Connect apps will have a path to adopt them.
we absolutely will not discontinue any Connect services until we have an equivalent or better option for partners to move to - and a viable path to get there.
Even if Forge implemented every existing Connect feature, it doesn’t cover the things that are only possible with Connect apps. We could not currently migrate fully to forge, because Connect (and therefore Forge) is still deficient in some API functionality that we have to use user API Tokens and direct REST to solve, the approach is not allowed in Forge for security.
It would be rather awkward to hear in a future Connect EOL announcement when there is no path to achieve functionality that was only viable for such externally hosted Connect apps.
This part really struck a nerve, because of the disconnect between delivering new Forge-only features and the lack of progress with regard to interoperability between Forge and Connect.
If you are really serious about being committed to supporting Connect builders, you would not release new Forge-only features if the path to adoption for Connect apps is not also immediately available.
Otherwise these just remain empty words, and only further contribute to the erosion of trust in the Atlassian Economy. Either you are committed to Connect, or you admit that it’s a deprecated platform. You can’t have your cake and eat it.
I think the point being made is that Connect will live on with certain features that are exclusive to it (like JWT, integration from third party platforms). I think my biggest worry would be to essentially cut off those features from the core products or penalize us on the Marketplace for using these features, calling Connect apps “insecure” etc.
I can live with the requirement to use Forge for certain apps like custom fields etc.
