What I learned at Gitlab that I don't want to forget
blog.boleary.dev
blog.boleary.dev
1. Nearly all features feel like they are half-baked. I do not even know where to start my enumeration of things that can be improved. The worst is maybe the search, followed by the wiki, followed by the issues.
2. It is more important for them to add more features then to improve existing features.
I think this is a direct consequence of the company politics as stated in the article. People may like working there and see it as gold standard. But from my point of view (as a paying customer) I’d like to see more people there work more on consolidation and refinement. Of course engineers prefer working on the cool new stuff. If there is no one who forces the focus on improving existing stuff, nothing gets better over time. At our company we have the same sickness, I wonder what others are doing differently that they don’t have the same problem. E.g. Apple.
On the other hand: there is not much of an alternative to gitlab. Atlassian self hosting is no more. Maybe Gitea + anything for CI/CD?
I think a lot of engineers miss the forest for the trees when it comes to stuff like this because what we value is good engineering and convince ourselves that such a thing is a prerequisite to the overall business success.
If you can’t convincingly argue why $work will have a positive ROI with engineer hourly rates then it won’t happen. How I’ve sold it in the past is having a bug bash where we groom and fix a couple hundred open issues and then use it as marketing fodder. It’s possible to make bugfixes as flashy as a new feature release if you sell it right.
I haven't used it yet, but something to look into.
Honestly, in larger organizations self-hosted GitLab is still one of the better options. It has lots of features and gets the basics right (though not all are excellent) - you get code review, CI/CD, even functionality for issue tracking, Wikis and much more.
You might use Jira or Atlassian for some of those, or maybe something more tailored and optimized for your workflows in smaller orgs, but GitLab still offers you options for most of the things you might want to do (even some monitoring and feature flags).
As long as you can keep up with resource requirements and updates, as well as configuration, you aren't likely to get fired for picking it and you'll be able to succeed and ship projects while using it, if self-hosting is a must/preferable.
> Atlassian self hosting is no more. Maybe Gitea + anything for CI/CD?
That said, for my personal needs, I migrated away from GitLab and run Gitea (for the code), Drone (for CI) and Sonatype Nexus (for container images, libraries etc.): https://blog.kronis.dev/articles/goodbye-gitlab-hello-gitea-...
In my case, dealing with updates is a bit easier now and breakages don't affect the whole solution (since previously I ran Omnibus GitLab for my own needs, which made launching it easier, at the expense of more coupling) and the total footprint would be much lower, if only Nexus wouldn't love to gobble that much RAM.
Sometimes integrating a bunch of specialized solutions is also a valid approach, especially because in my case container image cleanup in GitLab was a bit of a mess and while Nexus still has its problems (e.g. skipped deletion of old container images when they have multiple tags), it's a bit better and gives you more control.
Gitea is great, though and should work with most of the popular CI solutions as well - whether you want to run Drone, Jenkins, or anything else.
It's also quite frustrating to encounter the endless threads on years-long open issues with no solution in sight.
It had the upside of making some things more integrated. MR and issues and CI work much better together than GitHub or atlassian.
But it had the down side of expanding into feature categories half baked.
I only managed to stay for 4 years before I couldn't stand it. But I'm more an SRE than a feature developers.
I left for a job that has "burn tech debt" as a key goal.
But these values are nebulous concepts. The usually come from what the company already thinks it does rather than an aspiration. As an aspiration they would kind of suck.
My personal position is that you should think about the values of the organisation you work for and understand them, why they're there, where they came from. But don't internalise them, don't have corporate values in your heart. They are strictly limited and they aren't and can't be actually meaningful to the organisation itself because it's not a person. Someone there could still get you fired or do things that run against those values because they feel like it and the organisation as a whole will not protect you just because those values are written.
I've had a little interaction with GitLab and it wasn't good. Clearly some of the people I talked to were not following these values, they were not kind and understanding. But I know people who do work there and enjoy it. The company I work for talks about values a lot, and that's fine, I agree with the ideas behind them, mostly, but I think those talking about it have less control over the actual values and culture at the org than they think.
If you work for a big company, and the values were made at a corporate event, the values are a lie. They were made to sound good, but do not reflect the opinions of any one person. At best they represent some average of what the people most eager to declare generic human values at corporate events feel.
In either case, it’s just internal marketing. The true values of a workplace will be evident after a while, no matter what the values are declared as.
Especially, when a company has the value “transparency”, they usually don’t.
This kind of attitude - trying to extract the maximum benefit for yourself that the rules allow - is the reason why we end up with stupid policies like having to ask your manager, director and VP for $500 for a lab device, instead of just making an adult decision and buying it.
It feels so cynical and it poisons the well for the rest of us. It's the same energy as wanting to know an exact list of words that can get you in trouble with HR, as though you could then say whatever you wanted within those technical boundaries.
I would not want to work with many people in this thread.
> 1. Spend company money like it is your own money. No, really.
The author’s commentary:
> Instead of crafting the perfect policy to fit humans into, hire adults and expect them to act like it. If someone can't reign in expenses and stays at the Ritz-Carlton every night, the answer isn't to have a policy that prevents it, the answer is that person shouldn't work for your company. They aren't a good fit.
My take:
If I’d spend my own money to stay at the Ritz-Carlton (I like nice places, I have the money to pay for it, etc.), then why does this make me a bad fit for GitLab? If the company doesn’t want the Ritz-Carlton expensed then make a policy for it.
Either GitLab has “shadow” policies in place (annoying unwritten rules) or the author doesn’t understand the policy that GitLab has. I wondered if their Core Values had something about thriftiness which might disqualify the Ritz:
> GitLab's six core values are Collaboration, Results, Efficiency, Diversity, Inclusion & Belonging, Iteration, and Transparency, and together they spell the CREDIT we give each other by assuming good intent.
They go on to say:
> Every dollar we spend will have to be earned back. Be as frugal with company money as you are with your own. In saying this, we ask team members to weigh the cost of purchases against the value that they will bring to the company.
So they’re not really asking people to be as frugal with company money as they are their own. They’re asking people to be as cheap as GitLab wants them to be — without being specific about it.
Another example: “unlimited” time off is granted inequitably. Some teams and individuals are more pressured than others to stay within an unwritten amount of time off.
More to say, some studies [0] have shown that with Unlimited Paid Time Off "employees took less time off than previously, presumably leading to higher burnout rates".
[0] https://www.frontiersin.org/articles/10.3389/fpsyg.2022.8121...
Overall, I think Gitlab is building a culture that _to me personally_ looks weird, you know when you talk to someone and have a gut feeling that you can't trust them?
Did they fix location-based pay yet, many times discussed on HN in context of Gitlab? If not, any news on introducing Purchasing Power Parity?
It's the problem with these policies. They're written open-ended, but there is really a "right" answer and it's whatever the boss decides. Good boss? Maybe they stick up for you. Bad boss? You're going to regret that decision.
Imagine for a second that company is not reimbursing you for travel expenses, would you go for medium priced hotel or for 5* hotel with your own money for company business trip (not your leisure vacation)?
answer becomes obvious if you look at issue this way
That makes travel allowance a part of your paycheck, and all expected travel expense is loaded upfront in your salary.
then it becomes up to you - whether you go to luxury hotel for a business trip, or a medium range. Company becomes indifferent to your choices, and it becomes your choice how much to spend.
what would you choose this time?
same for business class tickets
The only situation in which it'd actually be obviously excessive is when the choice would be considered truly opulent and wildly expensive given all of the circumstances.
A 3 star to a 4 or 5 star is probably not so clear if all options are available.
If I was tasked with going somewhere on behalf of a company that I don't control, then I'd pick a place that would likely start at a cost of around $3-400 USD per night with no exceptional circumstances prevailing. All hotels basically suck, that range would let me not be concerned about where I'm staying, and do well the next day if I needed to.
Amazingly corporations decided budgeting it out would save them money. Even a 500/night hotel cap is better than the difference going in the employees pocket from the corporation vantage point.
Why are so many people so intent on assuming malicious intent on behalf of individuals and positive intent on behalf of organizations. Corporations have exactly one mandate: make as much money as possible. You are a line item, a cost. They will always squeeze you because it is the nature of the organization.
Yes. I would and have. Your personal means and current state in life shouldn't be part of company policy.
Policies like this are simply terrible for everyone. Those who would spend trip over some unspoken convention and upset others without knowing. Those who would not spend are left with a worse experience for no reason. No one has a clue about what is or isn't allowed, so everyone suffers.
It's nonsense.
They are looking for sensible people, not people who try to weasel every single definition and sentence. This applies to all communication, being able to understand others and read between the lines, and not try to one-up people by interpreting their words literally.
When the policy is "use common sense", that means "you don't have to go to a motel, you can expense breakfast, you can use the laundry if you're staying for a week or more".
Of course, people's perspectives vary. That's why you can talk and communicate to understand better. "Hey, I need to take a week long business trip, can I expense laundry?" It's not difficult.
> Bad boss?
If so, an open policy is the least of your problems.
In many cities (London, Dublin, etc) the price difference between shit tier and excellent hotels is marginal, and the excellent ones tend to be better located, leading to less travel/lost time.
The same can be applied to flights: taking the faster, business class ticket, direct flight with a flag carrier airline is often more economical in terms of "lost time" than using the budget airlines.
> > 1. Spend company money like it is your own money. No, really.
I normally maintain a ratio. When I was a junior, the rule I adhered to is spend $10 of company money like your would spend $1 of your own money. As I became more senior, I adjusted that ratio to $100:$1.
This keeps things in perspective but acknowledges that spending money generally makes money for a business.
So if I would pay $100 a night in a personal trip, I should be fine spending up to $10000/night of company money? I don't think that would fly at most companies.
Edit: Unless you're a C-level executive, I guess. But then you'd probably be fine spending 10k for a personal trip, too.
Will I book that executive class hotel within walking distance of my venues instead of La Quinta? Yes. Will I book that better plane ticket instead of cattle class? Yup. etc. Those are probably in the range of about a factor fo 4x-5x.
This also holds a little better if I talk about lab equipment. I have wasted a lot of time trying to do measurements with $100-ish pieces of equipment when I should have bought the $10K piece of lab-grade equipment that I then use for other projects as well. Now I try to optimize that better.
I certainly know that "reasonable expense" is a cultural issue, and strongly related to family wealth.
Someone coming from a working-class single parent family making $40K/year has a different idea of what "reasonable" means than someone from a professional class family with two working parents making $350K/year.
Just because these two employees now make the same money working the same job position for the same company doesn't mean they both automatically have the same idea of what's "reasonable."
For example, is it reasonable to get room service while at a hotel on company business? As someone from a $40K/year family, I never considered it until my boss mentioned it.
So yes, it's a cultural mismatch. But the conclusion shouldn't be that I wasn't a cultural fit (because I came from a working class culture), nor that I didn't have common sense, but rather that a company's specific corporate culture needs to be learned, and explicit communications may help.
Because you missed the crucial qualifying phrase, "every night".
The point of the policy wasn't that one should never, ever go out treat oneself to nice things. But that at some point, it get to be objectively (and grossly) inefficient and, to put it bluntly, reckless.
In fact, most people I know who in fact can afford to stay at the Ritz-Carlton every night (when traveling) don't, just as a matter of principle.
And yeah, they seem to also expect you to have a sense for what's called sound judgement, or shall we say business sense. That is, to know (at least approximately) where to draw the line, without an explicit policy for every last expense category.
And if you don't agree with this philosophy (or it isn't basically obvious to you) - then it would seem you probably aren't a good fit for GitLab.
(In the OP author's point of view, that is. The Ritz-Carlton analogy comes from them, not GitLab).
You seem to think more nights makes it a weaker case.
Travel is brutal, being away from home is brutal. The reason things like the Ritz and first class airline seats exist is that there are people who need to travel a lot and one way to keep them functioning at a high level despite the relentless pace is to provide superb comfort and service.
GitLab can decide to be that kind of company or not but when business travel becomes routine and frequent for one individual it’s dishonest to pretend they should scrimp like it’s their own much rarer leisure travel.
The ability to distinguish between the cost-value ratios of the two categories -- and to understand that yes, even for a large, successful company, unnecessary expenses of this sort do add up -- is precisely the sensibility that GitLab[*] is looking for.
[*] Again, a hypothetical GitLab - the Ritz-Carlton example comes from the blogpost author, and not from GitLab.
The only reason you tell someone you stayed at the Ritz is because you haven't stayed there often.
Gitlab is just asking people to use a little bit of common sense.
This is very clearly not what they’re asking people to do. That’s the problem.
2. This is way too similar to the Gitlab employee handbook, just glugging down the company kool-aid. Gitlab thinks they’ve figured out work culture despite producing an inferior product to competitors.
3. Writing everything down is exhausting. Asynchronous-only communication is exhausting. That’s a major reason why more companies don’t operate like Gitlab. Throwing face to face communication in the trash is the opposite of agile, might as well write letters to each other and deliver them on horseback.
The summary of this article is “Gitlab is so cheap they make employees write down everything they do and think rather than sparing any expense to facilitate regular in-person communication, and when they do send employees on-site they stiff them on traveling costs with a vague and passive aggressive travel policy that refuses to provide specific guidelines, which will lead to inevitable inequitable enforcement of unwritten rules among a workforce of diverse backgrounds, incomes, and cultural ideals surrounding money.
For example the Spending Company Money page [2] doesn't begin and end with the quote I shared - it just begins with it. It's the starting point, and in no way the articulation of the _entire_ policy. There are lots of guidelines past that to help folks out and understand what is reasonable...but the starting point is what is important. Trusting people to make good decisions for themselves and for the company is a very healthy thing, and I feel bad for folks who haven't gotten to experience that in work environments.
Thanks for all the commentary - it was a nice send off to make it to the top of this site for a time :)
[1] https://about.gitlab.com/handbook/values/
[2] https://about.gitlab.com/handbook/spending-company-money/
Edited for spacing and to say - one last time - GitLab employee here.
Which is a pity because their business sense is awful. The first major issue was the whole deleting inactive repositories debacle. They never officially announced it, but they never denied it, and were commenting on the outrage as if they were planning to go ahead with it. The whole way they handled that left a sour taste in my mouth.
So I casually started looking for somewhere else to host my code. And during this time, I suddenly realized something else: if you are not logged in, the "explore page" or whatever you want to call it has disappeared. I have no idea when this happened, but it shocked me.
It still exists (https://gitlab.com/explore) but there is no link to it anywhere. On github, they don't make it easy to find, but at least it is still there.
As a place that is pro open source, why would you go out of your way to hide all the open source code on your platform?
This utter disconnect between their values and their actions is very off putting.
I've moved all my stuff to codeberg.org and have been very happy since. At least they give a crap about open source software AND do something about it.
But gitlab say that they value open source. So that was the standard that I was holding them against. If a code forge claims to support and value open source software, why would they hide all the open source code on their platform? Why would they consider deleting stale open source repos?
I was just holding them up to their own standards, but I see they don't care about that.
It's similar to holding Google Chrome and Mozilla Firefox browsers to different standards. Google doesn't claim to give a shit about privacy, but Mozilla do. So when Firefox does some privacy blunder people get upset. But when Chrome does it, people don't care because they already had such low expectations of Google.
Some people try to live their lives by their principals, and not just "as a bonus".
If I would spend company money like I spend mine there would not be growth at all, just cost cutting and terrible amounts of not invented here syndrome.
I expensed one subway sandwich and stretched it three meals and felt guilty expensing a snack later in the evening. When they later clarified and said I had $70/day, I realized they have different life experiences for what a person spends. I also ate better than I ever had in my life.
Substitution: Bought something instead of spending time. Multiplier: Spent money to be more effective with my time.
Really think like an accountant. Realize your time costs the company $500-600 a DAY. And 1.5X to 2X a year than what they pay you. Save a week of time, $3000 not counting moving your ship date forward an inch. Something that increases your productivity by 5%? That's worth $10k a year.
Interestingly, I note a lot of people complaining about gitlab interviews but maybe that is the point. They tested you by disagreeing with something you wrote to see how you reacted. If you got defensive or didn't have the means to come back to their complaint, maybe you don't have the right culture!
There is a reason we like rules, policies and principles written down and enforced as written.
"There are no rules" With suggestions of "values" or whatever in place means you can't know in advance what an infraction is and you can be punished in an entirely unjust fashion with no recourse beyond appeal to those higher castes. This is a normal course of events in humans organised under a regime known as a dictatorship. Be it a smallish company or a country.
In a small organisation it really can work with little to no abuse. The larger the organisation gets the more horrible it generally becomes. At the country level it is essentially universally horrific for those who have to live subject to it. (Debate Singapore elsewhere, it's not interesting here).
The rule of law is a good thing. Being friends with the right people making what one person can do wildly different to another sucks very hard.
The flip side to note is that you can't make a bad employee (usually a manager) a good one by making policy pronouncements. You can't make (necessary) social norms accepted and socially enforced and adhered to by passing rules.
"We have no rules" presented as a positive for any community is a statement one should immediately be concerned about and consider its implications. It's up there with "There is no set amount of annual leave - take what you need" It just isn't at all likely to work in your favour unless you've decided you're short term and game it and even then it miraculously probably won't be approved or paid before you are cut for trying to actually take a reasonable amount.
Does anyone like HR and think of the HR dept. as a positive who doesn't work in HR? They're the pointy end of nastiness who are absolutely not on the side of anyone who is abused by someone senior. How's Harvey Weinstein's HR dept doing?
“I like spending all the money I have. Every dollar left unspent represents labor that has been given away without enjoying the fruits of that labor. If I die I won’t take any money with me.”
“You’re fired, you’re not a good fit.”
“You fired me for discriminatory purposes and used your intentionally vague travel and expense policy as an excuse, I’m suing you.”
I see 100 person startups with policies like this, not major corporations like Gitlab. Yikes.
I agree some amount of clear rules is good and keeps you from guessing, but the article says that the policy goes on to cover those, so what's the problem?
Not every employee-employer relationship is adversarial, no matter how much HN wants to believe it is. (Although, I grant you, beyond a certain size it tends to become adversarial.)
I agree that there a massive benefit to this and asynchronous communication can be more productive. However from a personal point of view, I’m dyslexic and find verbal communication much more efficient for me. Reading or writing can take longer than others expect. I also don’t always feel I can express myself correctly with written communication.
This is just a comment and unfortunately I don’t have a solution. I suppose this is more an ask to be mindful that others might excel in different ways.
the link is broken for me.
The link is https://blog.boleary.dev/gitlabs-values/#disagree-commit-and...
3 AM, New Year's Day. Awful timing. The author here talks about not being a fit- yeah that f---ery made it clear that company was not a fit for me. And I lost all respect for them, and have never seen a reason to offer them a 2nd chance, either as an employee or a customer.
[0] https://www.opm.gov/policy-data-oversight/pay-leave/federal-...
Maybe somewhere with a different calendar entrenched might not care so much about Jan 1 though.
https://www.opm.gov/policy-data-oversight/pay-leave/federal-...
It is a 2 week period where fuck all tends to get dome because somebody, somewhere is on PTO or otherwise unavailable.
No.
Being cognizant of other people's timezones/work schedules is an important part of remote work. If you call me at 3am, it'd better be urgent.
Agreed (and the GitLab person in a sibling thread said this as well), but I'm just saying, being a remote-first company should make you less likely to do this, not more. It'd be more understandable to me if they generally hired from a narrow window of timezones and recently expanded to hire from anywhere.
You're saying that I am generalizing too broadly from my experience working remotely? Just linking articles like this is kinda unhelpful, I'd prefer if you stated what your objection was, rather than leaving me to guess and to make your argument for you. That might not be as dramatic as linking an article like a mic drop but it's more productive and prevents misunderstandings.
What I'm saying is that in a remote company (distributed across many timezones anyway), you encounter this situation regularly. So you'd expect the culture of those companies to be prepared for this situation. If I'm missing something in that analysis, I'm happy to be corrected on it.
(I'd point out this was exactly the kind of confusion I was talking about. We had different ideas about what argument I was making, so when I tried to interpret what your criticism was, I guessed wrong.)
While it's possible that it could have been 3 AM on Jan 1 in Sydney and midday on Dec 31 in the United States, I think anyone at GitLab, including the recruiter who made the call, would agree that the timing of this call was a mistake - even if the recruiter was within their working hours.
Oops, we lost your data - but don't worry, the team here all agrees that it was a mistake.
On one of the technical tests, after getting through the meat of it, instead of having a nice chat etc the interviewer spent some time doing what I consider as exploring any possible area he could trip me up on and trying to make me feel uncomfortable.
I actually felt a little physically ill after that interview. Usually I quite enjoy interviewing. Someone having a bad day? Could have been, but there were two people on each call. I don't wish to work somewhere where people act like that.