Upcoming changes to user limits on Free tier of Gitlab SaaS
about.gitlab.com
about.gitlab.com
Some practical examples: - if I participate in a 48h game jam with my usual team, it would cost us 120$ to host our code on GitLab. For 2 days!! - the non-profit I develop for would need to pay 1680$ a year, despite all but 2 of our repos being FOSS. That is literally 3/4 of the budget we had last year!
Server load, bandwidth and disk space are the things that cost money and they are barely correlated with the number of users in all but a handful of textbook scenarios. Why does everyone insist on putting customers into price brackets based on such a useless metric?
If there is a curve "amount of money a company of a given size will pay", in certain applications, server load, bandwidth, and disk space utilization likely scales far below that trendline, whereas per-user tracks far closer. It also ensures there is effectively no upper bound on CLTV as a given customer scales out, even if net-usage only increases logarithmically/asymptotically.
To try and be even-handed and take a more generous interpretation as well, the former is also less likely to be cogently communicable to a customer, and may have painful UX when limits are hit.
Except I would say what instead of a certain applications this is true for the most of applications. If something doesn't uses tons of traffic/storage and is user facing then the resources spent per user are minuscule, you would spend more time and (even literal) energy to actually track and bill the user for the spent resources and somehow explain why some resources are billed for the insane rates (because you need to bill some thing for insane rates to make up $10 ARPU/month).
I had the same problem recently - I was musing about re-selling my monitoring solution (I'm using it myself, at like ~2% of the VPS I'm renting) and the first thing of course was the idea to base the price on the load each client would generate... and I pretty quickly come to a conclusion what I can't reasonably price things at $0.10 per unit, because the time I would need to spend on the setup is way, way more costly for me than that and that doesn't include billing hassles in any way.
So yes, billing per user is not only beneficial for the provider on the "user is paid wherever it uses the resource or not" but also allows the provider to not to spend time and money on actually implementing the way to bill the user for the resources.
and your costs track a certain number (e.g. number of checks/hosts)
why would you bill per user?
if you did all that and you have the right LTV and CAC numbers, at certain point you could basically run it on auto pilot
billing per user, but maybe you have 1 power user in a team, and a few extra accounts for non tech/backup/etc, which login once a month to check something
It isn't and the cost of developing something... is too much for a side gig with the intended purpose for the VPS to pay for itself.
> why would you bill per user?
In my situation I don't need to bill per user (and it would be counterproductive for me to do so in the long run), my point here is what my price calculation per resources spent doesn't correlate with my resources cost in any meaningful way, at least with anything with the word 'income' in it.
> and your costs track a certain number (e.g. number of checks/hosts)
In the end I just drew some arbitrary numbers for a checks/hosts numbers, slapped the price tag of a couple bottles of beer and called it a day. One person was happy with it and he uses only a couple of basic HTTP checks (so maybe around 0.002% of VPS? Who knows). Other one said it was too much for him. *shrug_emoji*
The same price for developers and testers who use it every day, but also the person who logs in twice a year to generate a report.
You end up with a product that's more difficult to use because costs end up influencing who can access it.
https://docs.gitlab.com/ee/subscriptions/self_managed/#billa...
If something is free, cheap, and relatively new, be prepared for disappointment.
It's like that for everything IT related. I've done part time IT service work for 20 years and in the last 5 years the cost of IT for small businesses has drastically increased. I'd estimate that things like email and network devices (management + updates) have increased by 5-10x for a lot of small businesses.
Family run businesses that are providing income for 1 or 2 people can't absorb those kinds of costs. Imagine if every appliance in your house all of a sudden became an annual subscription that's the same amount as the one time cost used to be. That's what small business IT is like now.
Normally I'd be optimistic and say all that means is a lot of opportunity for anyone that wants to start a business, but it's not. The high prices are being driven by consolidation and reduced competition in every area where there's some type of barrier to entry. Network devices are the best example. You can't decide to start a hardware manufacturing business without a ton of capital.
It's only going to get worse too. My prediction is that within 2 years Ubiquiti will go subscription based, within 5 years Windows will be subscription only for business use, and getting a perpetual version of MS Office will be nearly impossible within the next 3-4 years. I also think both Microsoft and Google will discontinue their non-profit / education offerings by the end of the decade because all the budget friendly options (aka competitors) will be completely eliminated by then.
Not going to happen because both of them are buying Mindshare: Microsoft is trying to salvage their staple of being a must-have requirement for office IT with Excel/other MS Office (including Office Online) dominance, while Google is trying to take all of that away and get people hooked on the power of collaborative document/spreadsheet editing, now with Notion-style features like inline people mentions.
In general it seems like Google is winning in lower education with the advent of chromebooks, MS is winning on the higher education front (Uni staff don't want to move away from Office), and Google Workspace is winning on the Startup front, but I doubt either offerings are going to charge schools for their basic offerings and will instead opt to throw more new features into their own separately-subscribed SKUs (eg. Google Voice having a different per-user subscription cost).
Hardware is interesting, there’s ongoing fees for the business if it’s connected hardware. However, almost all of them (b2c) charge a one time fee (IoT devices such as smart switches, security camera) and then charge a subscription for more advanced features or cloud storage (such as for the camera).
Also, it’s definitely a lot easier to produce and sell hardware products now compared to before!
For the Bay Area developer, sure.
For a South American developer, 10 dollars is about three meals.
Despite our low usage of resources and features that hadn't changed in a year, they suddenly wanted to charge us more than 10x a month, solely based on user count. If we dropped our user count below 7, we could still have the same pricing. It's absolutely absurd.
Zapier has similarly unfriendly user pricing; one of our team needed a few basic simple zaps for a workflow and since it's business process, we needed someone to have access to it too for management. That means we now qualified for the "team plan" which jumps the cost from $0 for the free plan (which already offered more features than needed) to minimum of $299/mo, with a whole ton of resources and features we didn't need, just to have an extra user.
For all that cloud pricing is difficult to predict, I'd much rather be charged by usage than arbitrary metrics like users (except where it makes sense).
For example: https://www.gitbook.com/pricing Team - $6.40 Per User Per Month *Starting At 5 Users*
Having said that, I think they're killing the golden goose.
At one point, Gitlab's competitive advantage was relatively generous free access to private Git repos. This generated a lot of goodwill in the days when GitHub didn't provide private repos even for personal use. I'm guessing that Gitlab now views its competitive advantage as the CI / DevOps experience. Still, it feels to me like they could limit that feature specifically rather than limiting even public projects to 5 users. Right now, GitHub's offering is suddenly looking attractive again for anyone who doesn't want to fill out a separate application (annually!) for each of their open source namespaces.
I'd think it's to Gitlab's advantage to keep these users. Right now, the current policy effectively chases them away. The only projects I imagine staying are those that really take advantage of the CI, but... is that really who Gitlab wants to limit their userbase to?
Since they're doing this, I assume they've done the analysis and determined this is no longer cost effective for them. But I'm surprised, because (a) they'll lose goodwill, and (b) what's their alternative plan for customer acquisition? I don't think Gitlab has grown to the point where they can afford to cut their growth, but I don't see how they're going to avoid that on the path they're going down.
I don't have any inside details, but I'd suspect that AWS also pays for PR, either through an internal team or external vendors. So "customer goodwill" is (probably) something they pay money for. Real dollars. This isn't free; you have to pay for it somehow. Whether it's dollars to fund a PR team (and the whole funnel of marketing/sales/etc.), or through free tiers designed to spread by word of mouth... you can't expect to acquire customers for free.
Bundling CI minutes on the paid plan doesn't change much, as plenty of customers will keep running their own CI runners for a variety of reasons.
Which means they're essentially very cheap users of Gitlab or would otherwise consider cheaper plans like 5$ a month
I think it might signal another thing too: "We're big enough, so we don't need that booster anymore".
Trello did that back then. They provided a middle tier called Gold which gave personal users all the goodies, sans the enterprise features. When they got big enough, they moved to strip it out and everybody (incl. me) were up in arms.
We've agreed in a trade-off, and then they still moved us gradually to enterprise. Now we're all using "Team of One" professional accounts with $99/yr. price tags. Trello is an enterprise platform and we're using it expensively just because we are a bunch of people who can use Kanban boards with katana-like precision for our personal projects and/or workflows.
GitLab might be doing this move to position themselves for bigger customers. Knowledgeable people can either pay, or use self-hosted versions they may think. Also, self hosted GitLab has evolved from a simple GitHub clone to a complete pipeline solution with extremely powerful functions. Even GitLab SaaS can talk with your in-house runners, so you can have an on-premise CI/CD pipeline with cloud frontend. It's kinda crazy.
Either like it or not, but you can't burn money like Microsoft + GitHub for free tier.
- Saved searches.
- Single board guests.
- Calendar, Dashboard and (rarely) Timeline views.
- Custom fields
- A lot of butler automations.
- A lot of power-ups, esp. Evernote, Dropbox, GitHub (it was limited before, so looks like they rolled gold into free somewhat).
Now there's no chance of that happening, as we aren't going to pay for both, and we aren't going to switch wholesale, so it's looking that I'm going to have to move everything to Github...
We were actually a Github paying customer and were going to pay to move to Gitlab, but given all those changes and rug-pulls, we decided to stay in Github. GH actions resulted to be enough for the CI/CD stuff we needed.
- $20/per user for most basic tier when I don't really need any of the features offered (except maybe "approval before merge required")
- $5/month for 10GB of storage ($0.50/GB) is steep. This stopped me from importing my large Git LFS project
- No way to have "view only" members for a repo or issue board without it contributing to your user number count and having to pay full price for them.
It's clear that I'm not that target market for them which is fine but its a bit of a shame as I do like gitlab.
The cheapest plan is 19$ month and it doesn't give us any extra benefits
and it's billed annually! Take it or leave it. For me it's 3 months timeline to leave Gitlab :)
I still like it, but not worthwhile for my professional projects and Github is a better homepage with all my OSS contributions displayed.
beside there's always a couple of people (non tech) who get added to the repo and barely login (but they want to keep the access), not worth paying 19$ for that
There really needs to be something like a "view only" or a role based cost per user. Most of our clients are just there for commenting, and only a handful of developers actually use the product to its full extent.
I know there was an issue open for discussion on their payment models, so I hope it is something they are continuing to look at. At 4x the cost of competitors that offer similar features, it is a very hard sell.
But I guess I don't use a lot of the features a business would need, so I'm only speaking for personal projects. I dunno, I've never actually looked at collaboration features for gitea.
I agree the implication feels a bit like don't bother complaining since you guys are an insignificant portion of the userbase. Cool, good to know I guess.
This has absolutely nothing to do with antitrust. The word antitrust has lost all meaning.
I'm fine with deciding to take away free stuff. The part that irks me is the lock in. When you create a lot of custom code to work on the free system they offer you, you are left to either pay or scramble to rebuild what you have somewhere else.
Agreed it’s a weird point to include in an announcement, especially when 100% of the people affected by the announcement are in the 1% of users.
"This impacts 1% of our users and saves us 10% annually on hosting costs. It's a no-brainer."
- It seems to me that we are promoting in IT a lot the idea of taking decisions based on data i.e data-driven companies.
- What does it mean to make decisions based on data? It seems to me that it looks exactly like this: we should look at what we want to change, look at the data and decide if the change will hurt/help/support a number of users. Do some math and then the result will be Yes/No for the decision.
- Now, of course, we could add "look at the data, but also think about second-order effects, about hidden relations ..." but that looks to me more like a No true Scotsman argument where only companies taking decisions that we like are really doing data-driven.
If we want data-driven decisions then why does it feel that the decision to impact only 1% of the users is wrong?
(please note: I am not saying this Gitlab decision is a good one. Probably it affects more than 1% of the users. I am mostly discussing the idea of the OP about "impacts 1% of our users")
I've always encouraged small non-commercial hobby projects to use GitLab. It just seemed like the right thing to do.
If you're familiar with Paradox Interactive Studios games (Crusader Kings, Hearts of Iron, Stellaris):
- Star Trek: New Horizons (127,000 current subscribers)
- SW:Fallen Republic (97,000 current subscribers)
- Kaiserreich (774,000 current subscribers)
are (/were) all hosted on GitLab.
A lot of the people who contribute to these projects only do so intermittently (so high member count to work done ratio) and are usually doing this in their formative years so would be likely to continue to use GitLab when they enter the working world.
There is absolutely no money available as game publishers prohibit it. And even if they tried to sneak in some Patreon donations there's no way that's going to cover $19/m/user!
Honestly, GitLab really seems to be falling far behind in their outreach. Their Open source programme is behind GitHub's and they don't have anything for Non-profits.
Doesn't that solve the issue for those projects?
Another problem is that game mods are often developed in private repos to prevent ideas and assets being "stolen" by other projects. There are no lawyers to protect your IP in the non-commercial world.
It's taken decades for the commercial world to even vaguely become OK with source available, let alone open source, we can't really expect teenagers and young adults who weren't trained as software developers to be comfortable with it.
Don't most game mods simply patch the executable if there's no mod framework available? I would think that the patch engine and specific patches aren't violating the original game's copyright, simply because they wouldn't contain any of the game's assets or code (aside from maybe memory addresses and executable names for hooking)
The point is generally irrelevant though, as a condition of uploading a mod to Steam or the game's mod system will be to give unconditional usage rights to the publisher which is incompatible with FOSS.
That being said, mod authors seem weirdly defensive and secretive about their code and ideas; I think it would be a hard sell to get most of them to use an open-source license even if it were possible.
My use-case for Gitlab is to support the internal tools and some bespoke development for a fan-run non-profit event. In a typical year we might have about $8,000 to run the whole event (ticket sales, less venue hire and other costs). Non-profit pricing is only available to registered charities in my country, so we're forced into paying full price for most of the SaaS services. That puts most of them out of our budget -- and it seems like Gitlab is joining that list.
So Gitlab are asking us to pony up $19 per user *per month* -- which is a cool $2300 a year. That's five times our current total IT spend (which benefits everyone), for a service which serves ten users. It's nearly a quarter of what we have to run the event.
Fair enough it unlocks extra features, but the only one I can see us using is Epics.
To compare prices -- Github is currently charging $48 per user, per year. That's $480 for the whole 10-person team. It's still going to be a tough sell for our Finance guy, but not as bad as $2300 (he'd ask if I hit my head if I asked for that).
The team picked Gitlab over Github because of the really good feature set on the free tier, and the availability of the bronze/starter tier for when we needed more. With Bronze going away last year, and now a 5-user cap, the message I'm getting from Gitlab is "we're a premium provider and we don't want your custom".
As @eslaught said -- it really feels like they're killing the golden goose.
If they added the option of buying more user seats for the Free tier and threw in a premium feature or two (Epics would win me over), I'd absolutely pitch that to my finance guy -- especially if the price was the same as Github.
Heck, I wouldn't even care if they cut the number of CI minutes... I can spin up a VPS if I need that.
Can an open core company survive a recession AND maintain their community/developer mind share?
They're still rapidly growing but they're loosing their edge to Github who since being acquired by Microsoft can afford to play the long game and give away more at a better price point.
Your second part of the statement seems more relevant.
The stack that they provide within the free edition alone is phenomenally consistent and capable compared to the collaboration tools that teams had available when they started - a mere seven years ago.
They are:
- OSI-approved open source license: All of the code you host in this GitLab group must be published under OSI-approved open source licenses
- Not seeking profit: Your organization must not seek to make a profit through services or by charging for higher tiers. Accepting donations to sustain your efforts is ok. Read more about this requirement here: https://about.gitlab.com/handbook/marketing/community-relati...
- Publicly visible: Your GitLab.com group or self-managed instance and your source code must be publicly visible and publicly available.
GitLab's discretion means that we review applicants to ensure they meet these criteria and do not violate our CoC. The renewal helps us to ensure that our program members continue to meet these criteria.
We've got a FOSS project, use hosting and support services to pay staff to write MIT and GPL code. But because our model is nearly identical to GitLab model we cannot get the GitLab FOSS blessing. Is this irony?
In this case it makes sense for them to apply these changes and other users will have to see whether it makes sense to pay for the service.
I use both near-daily for different clients and I have grown to prefer Gitlabs... the price is higher, but it might be worth it to some of the people in the daily HN github issue threads (eg https://news.ycombinator.com/item?id=30790593 )
I'm not saying it's unfair. But it makes me less interested in trying things just because they are "free".
Moral of the story is- use and buy open source! Not open core.
There is business risk in being trapped in non-free, but sometimes that is still the prudent option.
I like open source -- mostly because it can be debugged! -- but it's not the solution to every problem.
I guess we'll start planning the migration but it's really disappointing not to be able to pay a reasonable fee and then be kicked out on a 3-months notice.
I've seen complaints about this business model for a while on HN but never got the chance to understand Gitlab's reasoning for sticking to this pricing.
so you may find a better deal than "pay-per-seat gitlab.com" without the inconvenience of self hosting
Example
https://gitlabhost.com/pricing/single-tenant-gitlab-hosting/
MS ended up moving offerings from their paid tier to the free tier and Gitlab decided to offer this.
1. Daily subscription and auto-cancel. For example, you can subscribe for 3 days, then it'll automatically cancel your subscription after 3 days.
In Gitlab's case, they would enable hackathon teams to use paid plans for 3 days without overpaying for an entire month.
2. Customers can toggle specific features for very fine-grained pricing.
Essentially, it's to make SaaS pricing more on-demand.
Sure most of my projects uploaded are small toy projects, but this change completely turns me off to uploading any larger project that I plan to work on long term and hope for additional contributors.
Perhaps people need to start re-evaluating what "free" means in the context of a cloud offering. "Free" is a marketing strategy, not a charitable act.
In the case of Gitlab there's nothing preventing you from using the self-managed free tier which will effectively remove this limit, but guess what? Running it on your own will also have a cost for you. The same way it does for Gitlab when they give it for free through Gitlab.com
Gitlab used to have a basic free tier and a 5$ tier which was accessible
Since they will heavily restrict the free tier, and the 5$ plan was moved to 20$, many organizations have a serious financial decision to make
and all the fans of Gitlab, myself included, will have to consider options
and the other problem is pricing/per seat billing, is difficult for everyone, of course. But surely there's plenty of organizations which are probably a lot cheaper for Gitlab to keep as users than others depending on feature usage.
Guess what? They dropped support and the suggested replacement product costs thousands per year.
Am I entitled in that case too?
Small non-commercial projects get stepped on all the time despite contributing most of the important software in the world.
GitLab recently upped their pricing from $5/m to $20/m as well, so even paying members got a huge price increase.
Message received, I guess?
I'd rather self-host anyway. I use Gitea but it is kind of limited. Gitlab (self hosted) otoh is a horrendous resource pig
Is there something suitable for self hosting that's not such bloat? What does Gitlab bring that's difficult to self-host?
I wonder whether bug tracking can or should simply be folded into Git, so the tracker would be distributed just as the source control is.
This just adds a little impetus to that decision.
You'll need a mail service that GitLab can route transactional email through. All the difficulty is in configuring things post-installation, but it's well documented.
After being one of their most enthusiastic early adopters, Gitlab has by far been the most bait and switch SAAS product I have ever used.
Q. What does namespace in the context of user limits refer to?
A. In GitLab, a namespace is a unique name for a user, a group, or subgroup under which a project can be created. User limits are implemented at the top level group or personal namespace. For example: If a user has a group named top and two sub-groups under top named child1 and child2 with 4 users each, then the top namespace will have a total of 8 users, which is above the user limit of 5.
If I have Group A which has A/privateGit1.git with 4 users and A/privateGit2.git with 4 others different users then Group A has 8 users ?
Also it becomes more complex with guest users
We count the unique sum of users within a namespace which includes the users in the parent namespace (group), subgroups, and projects.
I created an MR to update the FAQ to make this more clear: https://gitlab.com/gitlab-com/www-gitlab-com/-/merge_request...
Thanks for the feedback.
So, all the open source projects who choose not to do that for various reasons, small teams and businesses, hobbyist projects like hackathons which generally have more than 5 people to a team.
You know, just off the top of my head.