Seeing GitHub acquired, I can’t help but imagine that Gitlab would go that route sooner or later. Their free offering is even more extensive than GitHub’s, and I don’t think there’s visibility into whether their revenue from paid plans is enough to offset that.
Since becoming backed by a major player is both a blessing (cash reserves to fuel the free offering!) and a curse (drive to increase shareholder value could go against longer-term community interests), I have mixed feelings about this trend.
Slightly tangentially, Gitlab has nearly caught up with GitHub and is arguably ahead in some ways feature-wise. I’m happy having built my small agency’s workflow around it.
I enjoy being able to drop a dotfile into repository root and have the product tested and deployed by CI to (in my case) AWS S3 or EC2, or get an email report if something prevented that from happening, after subsequent commit. On my roadmap is Terraform integration and having the automation provision all resources, but even in its current state I don’t think I can overstate the difference it makes, especially with a smaller team of engineers.
Some features of Gitlab at the moment are objectively inferior to GitHub’s implementation (simpler protected branch model with no way to mandate signed commits comes to mind), but many others appear stronger (issue tracker & boards, milestones, merge requests, the above-mentioned CI).
Feel free to give me a shout if you’re considering moving your business to Gitlab and I’ll share my experience.
> Seeing GitHub acquired, I’d speculate that Gitlab will go that route sooner or later.
Bingo. It's bizarre to hear people stampeding to GitLab when they're subject to the same financial and revenue pressures as GitHub or any other business. Somebody's paying for those servers and bandwidth and that somebody is expecting a return on their investment.
Whatever happens to Gitlab, you have the option of hosting the open source edition yourself. If they get acquired by Oracle in the morning, a fork could conceivably rise to take its place.
> The whole _point_ of Github was that you didn't have to host anything yourself. Just create an account and, bam, off you go.
So, just like gitlab.com?
> On top of that, even if you did want to host Gitlab yourself, it's "open core", not open source. You'd lose a ton of functionality.
That functionality is geared towards enterprise. If you are one, you should be doing your own risk evaluation. If not, you're probably fine with the open source edition.
I hear that again and again but GitLab is so massively complex now, that I don't think anybody without a company of at least 10 full time employees will be able to maintain and reasonably extend it.
The difference is that Gitlab has publicly stated, from the beginning, that the goal is an IPO:
> Stay independent so we can preserve our values. Since we took external investment we need a liquidity event. To stay independent we want that to be an IPO instead of being acquired.
This is such bull. They are backed by Google Ventures and Khosla Ventures who would vote with their preferential shares to sell GitLab for $7.5 Billion dollars without question. They're in the exact same bucket as GitHub except no where near as successful.
There’s no way Gitlab is turning down multiple billions of dollars like GitHub. They likely burn even more money than GitHub since their free offerings are bigger.
travis-ci has a pledge to keep their open source free tier always available. I don't know how much they stand by it but theyve said it on twitter and their plans page says "Always free for open source projects"
IMO being open-source isn’t such a minor deal, because it alleviates the worst-case scenario should Gitlab ever get acquired: if theoretical new management removes or changes price tiers, you can spin up your own on short notice. That management could reduce the effort put into maintaining their CE product or take it in an unfavorable direction and cause a fork, but the immediate damage is avoidable.
One possible caveat is performance. While I haven’t had to deal with this personally, some people reported that Gitlab can be a bit slow to run[0]. This doesn’t matter much if you use their hosted solution (if they have cash now, they can solve it by beefing up their hardware), but could result in higher than anticipated costs of running a self-hosted instance.
[0] Speaking of anecdotal evidence easy to find on HN and elsewhere. I’m yet to see any performance benchmarks comparing self-hosted versions of Gitlab and GitHub Enterprise (the latter starting at about $2500 per year) on similar hardware. It could well be that Gitlab is the fastest product for its maturity and feature set.
Since becoming backed by a major player is both a blessing (cash reserves to fuel the free offering!) and a curse (drive to increase shareholder value could go against longer-term community interests), I have mixed feelings about this trend.
Slightly tangentially, Gitlab has nearly caught up with GitHub and is arguably ahead in some ways feature-wise. I’m happy having built my small agency’s workflow around it.
I enjoy being able to drop a dotfile into repository root and have the product tested and deployed by CI to (in my case) AWS S3 or EC2, or get an email report if something prevented that from happening, after subsequent commit. On my roadmap is Terraform integration and having the automation provision all resources, but even in its current state I don’t think I can overstate the difference it makes, especially with a smaller team of engineers.
Some features of Gitlab at the moment are objectively inferior to GitHub’s implementation (simpler protected branch model with no way to mandate signed commits comes to mind), but many others appear stronger (issue tracker & boards, milestones, merge requests, the above-mentioned CI).
Feel free to give me a shout if you’re considering moving your business to Gitlab and I’ll share my experience.