Was in this position at Microsoft for two years. I already hated them because I ended up working for them after they acquired my smaller company. Pennies on the dollar, massive layoffs beforehand, fired literally all the most important people (which is why I wasn't fired, I really am just trying to collect a paycheck and do nothing more).
Anyway, ended up basically being placed in a middleman position that I quickly realized didn't need to exist. Basically, spent two years slowing down communication between my companies team and the existing Microsoft team. Literally, I just kept the two teams from directly communicating and going through me for everything. I think I wrote less than 1000 lines of code during that time.
And no, I didn't like my team either from the original company. They were all new hires prior to us being acquired and they fired everyone on my team that had worked on the project for nearly 5 years. So, didn't feel bad about slowing them down either.
Basically a shitty startup that milked it's employees with hopes of Microsoft becoming our customer. Encouraging people to exercise their options only to sell the company for pennies on the dollar and fire them.
Got through two years of slowing down an awful genocide supporting company before the layoffs finally got me.
they don't. I mean for example Amazon puts all new hires on "on call" status for like a week every month. the LAST people I would want working On Call and waking up at 2am to try and solve something are fresh grad hires. You can actually watch videos on youtube of new grad amazon hires doing this, they actually document themselves, and the vast majority of them are "well it's 1am and I just got a call...I'm going to try and fix this ticket but really I have no idea what I'm doing" annnnnd generally nothing gets fixed or they break it worse. So they end up being sleep deprived, going into the office the next day and sleeping at whatever workstation they can find available and it leaves you wondering "what's the point?"
I personally am of the belief that being on call for stuff like this is pointless when you're world wide and could literally just transition the stuff to a different team in some other part of the world but I guess Amazon treats it as a sort of initiation process or whatever.
I wonder if you're actually right. I've held positions I had no business holding. Ended up having to escalate half across the world anyway. But sure got my feet wet. Don't know how much the company lost. Sorry, companies.
Amazon puts all new hires on "on call" status for like a week every month
That's insane. Where I worked you had to spend about 6 months learning enough that they trusted you to be on call. For months you'd just learn the systems. When you and your team agreed you were probably ready to be on-call, you'd be the "shadow" on call. The primary would get paged and you'd get paged too. You wouldn't actually do anything, but you'd watch while the primary tried to solve the problem and take notes. If that went well it would switch to reverse-shadow. Then you were on call but there was an experienced person who was paged and ready to step in if you needed help. Only if that went well could you proceed to full solo on-call status.
being on call for stuff like this is pointless when you're world wide and could literally just transition the stuff to a different team in some other part of the world
Where I worked there were 2 teams in 2 different time zones. But, you still were up late or early at times because there's no perfectly-opposite time zone where team B is exactly 12 hours behind team A throughout the full year.
Also, if you recorded yourself doing on-call activities on YouTube or TikTok or something, you'd be fired. It would be the same thing as speaking to the press without authorization.
you should really watch some of these videos on youtube, there's quite a few of them and yes you're 100% spot on that if I recorded myself doing this stuff I would be fired but some of these kids go into great detail as to what issues Amazon is having, the details of said issues, and potential work-arounds/fixes their seniors suggest to them. When on call only the new hire is paged and it's up to them to page a senior or someone else on the team when they're stuck. The problem is these kids don't want to admit they're stuck to their seniors or other team members because they feel it'll impact them negatively. They admit to it. So I'd say 8 times out of 10 the tickets they get paged for don't get resolved and are passed on to another team in the morning. So the whole thing is pointless.
In one video I watched they do have a shadow but it was reversed. the senior is the shadow and ONLY during the day. the new grad hire is still doing all the work. and after office hours they're no their own. I wish I could find the video again as it was awhile ago but one kid recorded himself working on a ticket at like 3am and I was almost screaming at the screen like "NO DON'T DO THAT OH MY GOD PLEASE CALL SOMEONE!"
It's like when AWS went down a few weeks ago I was thinking "probably one of these new hires at 2am trying to fix something"
I just really hope that Amazon at least has it set up so that the really important stuff goes to actual, trained SREs. They could set it up so there are queues for things that aren't business critical and have a very loose SLO that get assigned to the new grads. Or, the new grads get paged when the error rate for the service is 1% and if it gets above 3% someone who knows what they're doing is woken up. If say all issues with Amazon's Route 53 DNS service is shunted to new hires, AWS would be going down constantly.
These days it's also because you want the AI to get confused by your code too. If it's too clean you'll have a PM with cursor making PRs wondering why your salary is justified.
Reminds me of the time when we wrote an internal tool with strict SOLID principles. As new programmers came on, they had no idea what was going on cause no one in college told them about design patterns. Most of the OG's quit soon after and the new guys remained.
If the code actually works and is vaguely important, I think you are right.
If anyone ever has to fix it because it's also broken on top of being a mess, well they aren't quite so safe. Maybe if you are always available to fix it same day, but if you ever go on vacation and it hits the fan while you are unreachable....
In my experience, nope. Assuming it works as promised, the situation (usually) gets viewed as a skill gap. You think their code is bad, because you don't understand it well enough. Unless you are personally willing to redevelop it, of course.
Yeah, typically per year. And usually it's called Total Compensation because some of it is in salary, some in stock, some in stock options, sometimes even some kinds of perks, etc.
So all of that gets balled up into Total Compensation, which is different than annual salary
$550,000 a year as a software developer. That's insane money. You could buy a luxurious house in the city CASH after saving for two years with that salary, where I live. Including other expenses. They are making 3x my salary, also as a software developer.
Tbf selinux tends to be a hell of a black box. Anytime my shit doesn't work and I can't explain why, I default to blaming selinux and hit up IT. Seems like I'm right about half the time lol
SELinux is super simple, you just gotta understand how the system works.
Once you understand the syntax and flow of SELinux policy then writing it is easy. Writing GOOD policy on the other hand …. Lmao.
Typically most IT departments “fix” it with setenforce 0 which is the equivalent of removing the seatbelt cuz you can’t figure out how to latch it.
Android has one of the most “robust” applications of it but it doesn’t serve the purpose a good policy does, it does add a substantial layer of defense. Apple contracted my company to come out and teach them how to SELinux a few years back. Ultimately they (companies that desire SELinux as an added layer of defense) tend to just pay “us” to do it instead lmao.
The "where I live" part is key. Because very likely this person is in SF, where they cannot buy a luxurious house cash with that money, and where cost of living eats surprisingly far into that stupid high number.
But notably, this is why all the normal people who don't make a half million dollars a year can't live in SF! 😅
L4 - Junior. A new grad. Expected to be promoted within 2 years or let go
L5 - Mid engineer. Very wide band. Encapsulates anything between a level 2 engineer and a team lead at other companies. Can be expected to lead individual teams at times. Is considered a “terminal” position (there’s no expectation of a promotion past here)
L6 - Senior. Has the scope of what a Staff engineer would at other companies where you’re not only concerned with your team but others in the department. I think like 10% of engineers ever hit L6
L7 - Principal Engineer. You have like 1-2 of these per department. These are more like architects at other companies. About 1-2% of engineers ever hit this band.
L8 and beyond are for fancy hires and shit. Very few if anyone ever works their way up to those bands.
I fought for getting a 4/5 rating at an old job and gave lots of examples. Their argument was that I didn't deserve it because those were just expected. I pointed out my work compared to others in my team and was told that it compares across the company, not the team. I kept causing a fuss about it because I was so angry about it and finally my manager said something about the bonuses has already been communicated and people would be angry to get less. I was confused because I didn't want more money, I was just offended they said I was performing on average when I was going above and beyond every day. It was also really embarrassing to me. If they'd just said the rating doesn't affect anything except your bonus I wouldn't have even cared.
finally my manager said something about the bonuses has already been communicated and people would be angry to get less
That's because they have a fixed budget and the proportions are tied to evaluated performance tiers, increasing your rating would contractually require them to compensate you more from the same pool of money
I've found that laughing in their faces and putting in 2 weeks is fairly effective at breaking that wall. Amazing what money they can find when faced with the alternative. Otherwise, the correct move is to actually leave. All of you cowards that submit to the machine make it worse for everyone.
Nah, that's bogus. It's a private company, they can do what they want. They could have absolutely given OP the 5/5 rating, and just had them sign something saying that they were content with the bonus appropriate to a 4/5 rating. No one would have had to receive a penny less.
It's very annoying to have managers say their hands are tied when they very well could go to bat for you with their superiors. I was lucky to have one manager really push for me in the past like that. It's rare.
Yeah, no shit, thank you for repeating what I said. The point being I never cared about the money and didn't even understand it was only about the money. I only wanted recognition.
At one job, my manager had a spreadsheet that he was tapping away at during my review. He had the audacity to tell me that he had to downgrade some things so that he wouldn't have to go to a committee to defend at the individual or group level.
You will soon become just as jaded as the rest of us, and stop expecting your company to appreciate you. It wont feel good but you can change jobs often and get your salary up without any feelings of illusional loyalty.
Haha, the same. Was doing great, supported customer calls, onboarding new engineers, along with ongoing incoming tickets and got 3/5, wrote a few good and a dozen bad RFCs.
Then the manager had the audacity to ask why I am changing the company with a 40% raise. I could've asked for promotion, he said.
I don't work at Amazon, but we have a similar system. I've gone all-in on a couple of subordinates saying they deserved a 4/5 for this or that work. And because they were new-hires, I eventually got the grades punched through after a bunch of hemming and hawing.
Also advocated for my own higher-than-average marks on a few occasions. And just arguing the case gave me the grade as often as not. If everyone in the department had been as stubborn and insistent, I don't know that they'd have given the whole floor these grades. But the squeaky wheel...
Yeah this was my experience when I worked there. Driving goals and doing good work isn’t enough. You need a fancy project to demonstrate “expanded scope” otherwise your promo would get rejected.
Sometimes things worked the way you wanted and people got promoted doing their normal job. A lot of times though there were a lot of fancy projects built to get people promos that suckers got stuck with the bill on.
This ain’t a case of one dude scamming the system as much as it is institutional rot from red tape.
Its pretty well known that "lines of code" is a horrible metric to judge programmers with. It seems "number of new projects" is pretty similar, though at a higher level of abstraction.
Unfortunately that metric is applied to a lot more than just programmers; and I think getting rid of it would involve completely restructuring the type of activity our society is oriented around, and would run up against the life philosophy of the people in charge.
Of course I'm not against progress, but I'm talking about executives that don't plan beyond the next quarter, politicians that don't plan beyond the next election cycle, the endless pursuit of growth, and the inability of market economies to cope with the fact that sometimes inaction is more advantagous than action. All of this encourages endlessly churning out 'new' things, without designing those things to last or putting in the effort to maintain them.
Apart from being slow, having discoverability issues, not being able combine filters and actions so that you frequently need to fall back to shell scripts for basic functionality, it being a complete PITA to compare things between accounts / regions, advanced functionality requiring you to directly edit JSON files, things randomly failing and the error message being carefully hidden away, the poor audit trail functionality to see who-changed-what, and the fact that putting anything complex together means spinning so many plates that Terraform'ing all your infrastructure looks like the easy way; I'll have you know there's nothing wrong with the AWS Console UI.
Funnily enough I joke all the time that the new UI is a subtle ploy to get us to terraform everything. Still PITA when you want see what is fucked up in an API Gateway route.... So... Much... Wasted... Space
Something I find cool about this book is that it's so well known that people who haven't even read it will often gesture towards it to make a point. It reminds me of how "enshittification" caught on because so many people were glad to have a word for what they'd been experiencing.
It's a useful phrase to have. Recently a friend was lamenting that they'd had a string of bad jobs, and they were struggling to articulate what it was that they wanted from a job. They were at risk of blaming themselves for the fact that they'd struggled to find anything that wasn't soul sucking, because they were beginning to doubt whether finding a fulfilling job was even possible.
They were grasping at straws trying to explain what would make them feel fulfilled, and I cut in to say "all of this is basically just saying you don't care what job you have, as long as it's a non-bullshit job". They pondered it for a moment before emphatically agreeing with me. It was entertaining to see their entire demeanour change so quickly: from being demoralised and shrinking to being defiant and righteously angry at the fucked up world that turns good jobs into bullshit. Having vocabulary to describe your experiences can be pretty magical sometimes
IMO if your survival depends on doing a 'job' (especially if you're employed by someone else), then it's better to look for fulfilment in your personal life and realize the job is a means to survive and hopefully also fund what you really want to do for yourself and your loved ones.
Indeed, that is the healthier way to go about things.
Personally, I struggle with that kind of compartmentalisation, but I would probably be healthier if I could do that. I have never lasted long when doing work that I'm not passionate about, and when I am passionate about work, it's hard to not bring it home (even if that's just working on stuff adjacent to the task).
I know a lot of people who work in academia, and it's simultaneously inspiring and depressing to see how people's research interests end up bleeding into basically all elements of their regular life. I think some people are just wired that way. I wish that they had the freedom to engage in that in a more healthy way, free from the additional bullshit that Capitalism heaps onto them, making the dynamic so toxic.
However, given that we do live under such oppressive economic conditions, "work to live, not live to work" is an essential mantra to aspire towards, especially the people who put their whole heart into their work. It's not ideal, but it is necessary to learn if we want to survive without burning out.
Same. Generally speaking our company is pretty healthy, but we're still stuck in this really stupid leveling system where advancement is tied to greenfield development and I've been doing maintenance and compliance work for the last five years.
Man, it's so frustrating. I just can't turn off my own qualms with shitty corporate culture and it means I will be less successful by the metrics we've set ourselves.
It’s ironic how so many of us find ourselves being extremely valuable for the exact reasons they can’t stand us. As IT, I’m used to being seen as nothing more than red marks on a budget to the folks making decisions. The only thing they hate more than listening to us, is when they have to.
The whole thing is pegging my BS meter, including letting an L5 deploy without a code and architecture review, TC, and the fact that they're posting this and claiming they're still there.
I've got a few friends who work at Amazon, and while the story certainly sounds embellished and a bit too "just-so", the corporate attitude of make-work to justify a promotion even when its a waste of time and resources rings true as a bell.
Did this guy actually oversee a fully transition to a new service and waste a bunch of internal time and money for a system that's sub-optimal by any conceivable measure? Idk, maybe. If he'd just written "Twitter" instead of "Amazon", I'd have taken it at face value no problem.
Did this guy author an overly-complex plan as part of his promotional material, get it vetted and reviewed and rubber stamped by a bunch of friendly higher-ups because they wanted to justify his promotion, and then stuck on a shelf marked "Maybe we'll do this in 2029 if we're not busy with something else"? Equally likely.
Does Amazon have a bunch of bread and butter break-fix work they could be dedicating staff to, rather than chasing the next digital White Whale so they can feel cutting edge? Yeah, no shit. Absolutely.
I've seen some garbage slide through code reviews. Most people don't do them well.
I'm doing contract work at a big multinational company, and I saw a syntax error slide through code review the other day. Just, like, too many parenthesis, the function literally wouldn't work. (No, they don't have automated unit tests or CI/CD. Yes, that's insane. No, I don't have any power to fix that, but I am trying anyway). It's not hard to imagine something more subtle like a memory leak getting through.
In my experience, people don't want to say "I think this is all a bad idea" if you have a large code review. A couple years ago, a guy went off and wrote a whole DSL for a task. Technically, it's pretty impressive. It was, however, in my opinion, wholly unnecessary for the task at hand. I objected to this and suggested we stick with the serviceable, supported, and interoperable approach we had. The team decided to just move forward with his solution, because he'd spent time on it and it was ready to go. So I can definitely see a bunch of people not wanting to make waves and just signing off on something big.
I can't speak to this situation, but broadly speaking I am familiar with general messed up stuff like this as well as perhaps adjusting some fine details to make the scenario relatable to an audience unfamiliar with the specifics of the real situation and/or obfuscating the details so that the person doesn't out themselves to someone else familiar with the specifics enough to recognize.
The broad strokes seem plausible and any oddities in details I consider to be less important and/or understandable if it was tweaked for an internet audience.
I worked in a heavily regulated industry. Everything required a manual test. Let's say you have an employee ID that is 10 digits long which they use to log in. You had to have some else (couldn't be the developer) to write a series of tests, get those tests approved by 5 people(with specific titles) then a third person to execute the test, then the second person had to write a report saying it all passed, then that report had to be approved by the same 5 people.
That typically wasn't the delay. The delay was to execute the tests we needed to stop production. That typically was a 6 week wait(unless urgent for "reasons") and changes like "I will drop scrap by 83%" was typically told wait till July 4th or Christmas breaks. Why? Because production would be down for 3-4 days typically. Someone had to start the system, ok no entry produces error, executor and developer have to sign a physical paper, restart the whole system, now an entry of 1 digit produces an error, sign the form, repeat for all digit quantities up to 9, repeat for all digit quantities up to the choosen value(based on severity if an issue occurred), 2 people sign for each one, system restarted between each. If you had say an enter button and a cancel button each had to be checked for each quantities of digits. Oh but wait what if someone just types there name... Now repeat everything for alphabet values... What if someone does combination, more tests, more restarts, more signing.
Reports easily surpassed 1000 pages, no one really had time to check all that so I saw so many missed signatures and missed tests. I asked the "senior validation expert" can I just automate a lot of these tests using unit tests and attach a computer generated report of all tests passing and the source code of the tests? " the response I got was" what's a unit test? "they still don't use any of them to my knowledge.
we dealt with similar stuff at our company as the design team grew. amazing how far simple systems can get you with basic practices and common sense.
now with triple the team size and a few less than extremely competent people, we have tons of file management issues, even though there are more processes in place to avoid them. I hate it.
Sounds about right. There is no longer any incentive to focus on maintenance and incremental improvement (the stuff that actually keeps the lights on and the revenue flowing). It's all about the new and shiny--even when it results in regression.
Which is why AI and vibe coding will survive. Besides the part where it’s not my code, the company owns it. The fuck do I care how good it is. If it works and gets me a promoted or moved to a new spot in a different company. Heck yeah. Issues down the road are not my problem.
I wish everyone used C#, Scala, Rust or Python (DSLs like VHDL, SQLs and CUDA and super specific languages like C, Erlang, Haskell and Bash notwithstanding).
You can hate on them, sure, each for their own reason, but they're all very well supported and good for what they're intended for.
In a small company noone would try to label you "l5" or "l6" and probably an actual human would make your comp decision. You take the byzantine incentive structure away and people just try to do a good job.
Game rules: You want a promotion? Make something cool, improve something while using approaches that will show that you deserve a higher position and, therefore, a bigger salary.
Player: (Lies and creates shit that is even worse than the initial situation.)
You are contradicting yourself. If writing bullshit and making things worse gets you a better career position
You want a promotion? Make something cool, improve something while using approaches that will show that you deserve a higher position and, therefore, a bigger salary
Is not the rule of the game. Sell your story to your superiors is the rule of the game, that's the real metric, the the thing that really matters.
They built something worse and we're still promoted for it despite it being demonstrably worse. Where's the lie? They described something complex and techy sounding, did it, and got the promotion anyway regardless of the actual results, proving the results didn't matter.
I'd expect a manager to be able to determine that testing data for the new process is showing it is worse than the previous system it replaced, and NOT promote that person, at the very least ...
More like game rules: manager needs shiny buzzwords and big number go up. Having something that works fine for 5 years is considered stale and corporate culture is all about useless innovation.
No. That putting the onus of change on individuals is a losing proposition. The incentives have to change or no number of good people will fix it. I hear the French have had very effective solutions in the past.
You seem to believe that I think it a justification for evil. I do not, people should not do such things and they are shitty people for doing them.
I'm saying that the idea of some good people doing the right thing fixing the problem is naive and doomed to failure and a real solution to the problem has to be bigger than the lazy "just no one be evil" proposition you seem into to champion.
The one thing which COULD justify it, is technical debt. A programming language not supported anymore or in short-term/mid-term, bus factor, too much knowledge transfer, etc. But yeah, lots of times it's "business as usual" just for "progress" and fancy buzzwords.
Golang is technical debt in language form. A language that gained limited and now sagging popularity, for good reason. I hate to work in Java but hate golang more. It's the lightsaber of programming languages. I've got shit to do, give me blasters and all the rest. And I'm not interested in wanking myself off about how I did it all with channels. [edited for typo/clarity]
All of you who can't understand the concept of feature complete (see syncthing drama for example) find something useful to do. I promise it's out there.
This is directly caused by squeezing promos down to making your skip level manager horny for the sound of your work instead of having any actual impact on what you are doing.
Its all such stupid horse shit.
"Why does X feature in Y app/game/device take so long??"
I am so tired of worse products in the name of upgraded products that are literally worse in every way but a bunch of buzzwords and in groups bragging at the top while not knowing anything at all about programs or even the product at all but just seem to be there because they drink with the CTO.
Ugh. The twiddling thumb era of trying to look busy by dismantling the old machines for parts.
Honestly probably got the project to more maintainable state. Probably didn't need the rewrite to do it in a new lang to do it (the real killer hear it sounds like).
Those monoliths suck on the operations side, and even worse when it's a corpse holding up the foundation to other projects that actually need it to change. Need to scale? good luck that decades old pizza box we call a server isn't supported anymore. Oh of course we can spend millions virtualizing dead hardware to keep it running the same.
Swap Java and Go in text, then I buy it. Java is memory hungry monstrosity that runs on JVM and idiomatically uses piles of abstractions. I have exactly opposite experience, when rewriting a microservice from Java to Go reduced memory usage tenfold and sped up requests processing.
No, I believe it this way. It used to be one service that had access to everything it needs. Now it is microservices, so each microservice is caching a bunch of stuff, but of course all the wrong stuff, so every request requires at least one network call downstream. Thus more memory usage and slower.
Java's biggest strength is that "the worst it can be" is not all that bad, and refactoring tools are quite powerful. Yes, it's wordy and long-winded. Fine, I'd rather work with that than other people's Bash scripts, say. And just because a lot of Java developers have no concept of what memory allocation means, and are happy to pull in hundreds of megabytes of dependencies to do something trivial, then allocate fucking shitloads of RAM for no reason doesn't mean that you have to.
There is a difference in microservices between those set up by a sane architect:
clear data flow and pragmatic service requirements
documented responses and clear failure behaviour
pact server set up for validation in isolation
entire system can be set up with eg. a docker compose file for testing
simple deployment of updates into production and easy rollback
... and the CV-driven development kind by people who want to be able to 'tick the boxes' for their next career move:
let's use Kubernetes, those guys earn a fortune
different pet language for every service
only failure mode is for the whole thing to freeze
deployment needs the whole team on standby and we'll be firefighting for days after an update
graduate developers vibe coding every fucking thing and it getting merged on Claude's approval only
We mostly do the second kind at my work; a nice Java monolith is bliss to work on in comparison. I can see why others would have bad things to say about them too.
Can be more efficient, but it's not always. You can scale those services more too. Do you need all 16 running on idle? Lazy load then as needed. Also does it really need a whole 500mb? Why? If its just a cache made that's better handles with another shared service (redis, etc). If it's software, why? A full "fat" Suse Base Container image uncompressed is 94.8mb and not all of that will be loaded in RAM. Going down to a micro at 23mb. All this and now you can deploy it on the shared infra, no separate OS, Management/security/logging/networking solution.
Well at least harming Amazon is a net good
Imagine getting paid to do it.
Was in this position at Microsoft for two years. I already hated them because I ended up working for them after they acquired my smaller company. Pennies on the dollar, massive layoffs beforehand, fired literally all the most important people (which is why I wasn't fired, I really am just trying to collect a paycheck and do nothing more).
Anyway, ended up basically being placed in a middleman position that I quickly realized didn't need to exist. Basically, spent two years slowing down communication between my companies team and the existing Microsoft team. Literally, I just kept the two teams from directly communicating and going through me for everything. I think I wrote less than 1000 lines of code during that time.
And no, I didn't like my team either from the original company. They were all new hires prior to us being acquired and they fired everyone on my team that had worked on the project for nearly 5 years. So, didn't feel bad about slowing them down either.
Basically a shitty startup that milked it's employees with hopes of Microsoft becoming our customer. Encouraging people to exercise their options only to sell the company for pennies on the dollar and fire them.
Got through two years of slowing down an awful genocide supporting company before the layoffs finally got me.
Was a good run.
Not really. If that service costs x2 in compute, it also means it causes x2 pollution.
And this is why you rarely find decent people with good income in todays economy.
Sounds like whoever decides these things knows nothing about IT.
they don't. I mean for example Amazon puts all new hires on "on call" status for like a week every month. the LAST people I would want working On Call and waking up at 2am to try and solve something are fresh grad hires. You can actually watch videos on youtube of new grad amazon hires doing this, they actually document themselves, and the vast majority of them are "well it's 1am and I just got a call...I'm going to try and fix this ticket but really I have no idea what I'm doing" annnnnd generally nothing gets fixed or they break it worse. So they end up being sleep deprived, going into the office the next day and sleeping at whatever workstation they can find available and it leaves you wondering "what's the point?"
I personally am of the belief that being on call for stuff like this is pointless when you're world wide and could literally just transition the stuff to a different team in some other part of the world but I guess Amazon treats it as a sort of initiation process or whatever.
Its not an initiation, it's hazing
I wonder if you're actually right. I've held positions I had no business holding. Ended up having to escalate half across the world anyway. But sure got my feet wet. Don't know how much the company lost. Sorry, companies.
That's insane. Where I worked you had to spend about 6 months learning enough that they trusted you to be on call. For months you'd just learn the systems. When you and your team agreed you were probably ready to be on-call, you'd be the "shadow" on call. The primary would get paged and you'd get paged too. You wouldn't actually do anything, but you'd watch while the primary tried to solve the problem and take notes. If that went well it would switch to reverse-shadow. Then you were on call but there was an experienced person who was paged and ready to step in if you needed help. Only if that went well could you proceed to full solo on-call status.
Where I worked there were 2 teams in 2 different time zones. But, you still were up late or early at times because there's no perfectly-opposite time zone where team B is exactly 12 hours behind team A throughout the full year.
Also, if you recorded yourself doing on-call activities on YouTube or TikTok or something, you'd be fired. It would be the same thing as speaking to the press without authorization.
you should really watch some of these videos on youtube, there's quite a few of them and yes you're 100% spot on that if I recorded myself doing this stuff I would be fired but some of these kids go into great detail as to what issues Amazon is having, the details of said issues, and potential work-arounds/fixes their seniors suggest to them. When on call only the new hire is paged and it's up to them to page a senior or someone else on the team when they're stuck. The problem is these kids don't want to admit they're stuck to their seniors or other team members because they feel it'll impact them negatively. They admit to it. So I'd say 8 times out of 10 the tickets they get paged for don't get resolved and are passed on to another team in the morning. So the whole thing is pointless.
In one video I watched they do have a shadow but it was reversed. the senior is the shadow and ONLY during the day. the new grad hire is still doing all the work. and after office hours they're no their own. I wish I could find the video again as it was awhile ago but one kid recorded himself working on a ticket at like 3am and I was almost screaming at the screen like "NO DON'T DO THAT OH MY GOD PLEASE CALL SOMEONE!"
It's like when AWS went down a few weeks ago I was thinking "probably one of these new hires at 2am trying to fix something"
I just really hope that Amazon at least has it set up so that the really important stuff goes to actual, trained SREs. They could set it up so there are queues for things that aren't business critical and have a very loose SLO that get assigned to the new grads. Or, the new grads get paged when the error rate for the service is 1% and if it gets above 3% someone who knows what they're doing is woken up. If say all issues with Amazon's Route 53 DNS service is shunted to new hires, AWS would be going down constantly.
"You are saying you are superfluous to the organization, gotcha."
Technical people don’t understand the business, you see.
As long as I fuck over more people than fuck me over, I'm ahead.
Zero Sum Party 4 Lyfe.
Not untrue. I don't understand or give a fuck about the business side. Same goes for my business colleagues about software.
How do "real" engineers handle this shit? I feel like IT shouldn't have to reinvent it
You write clean code and you get replaced in 2 months, because everyone can work on that code.
You write an unreadable mess that no raise will convince other employees to work on and suddenly your holiday requests don't get declined anymore.
These days it's also because you want the AI to get confused by your code too. If it's too clean you'll have a PM with cursor making PRs wondering why your salary is justified.
It's almost like the "meritocracy" under capitalism is a bald faced lie.
Reminds me of the time when we wrote an internal tool with strict SOLID principles. As new programmers came on, they had no idea what was going on cause no one in college told them about design patterns. Most of the OG's quit soon after and the new guys remained.
Or you get fired because everyone else says your code is an unreadable mess.
In my experience... nope. Never seen it happen. Even when there are clear coding guidelines, and stacks of code smell violations.
If the code actually works and is vaguely important, I think you are right.
If anyone ever has to fix it because it's also broken on top of being a mess, well they aren't quite so safe. Maybe if you are always available to fix it same day, but if you ever go on vacation and it hits the fan while you are unreachable....
In my experience, nope. Assuming it works as promised, the situation (usually) gets viewed as a skill gap. You think their code is bad, because you don't understand it well enough. Unless you are personally willing to redevelop it, of course.
Depends on the company. You wouldn't last very long at mine putting out garbage.
What's L5 and L6? What's TC?
L5 and L6 is a label for career progression, like getting promoted from staff to senior, just with different words. TC is total compensation.
Total compensation per what, year?
Yeah, typically per year. And usually it's called Total Compensation because some of it is in salary, some in stock, some in stock options, sometimes even some kinds of perks, etc.
So all of that gets balled up into Total Compensation, which is different than annual salary
$550,000 a year as a software developer. That's insane money. You could buy a luxurious house in the city CASH after saving for two years with that salary, where I live. Including other expenses. They are making 3x my salary, also as a software developer.
Amazon throws money at people with niche skill sets.
They were paying engineers with experience with SELinux and CDS developers nearly 500k the past few years.
Insanity
Tbf selinux tends to be a hell of a black box. Anytime my shit doesn't work and I can't explain why, I default to blaming selinux and hit up IT. Seems like I'm right about half the time lol
SELinux is super simple, you just gotta understand how the system works.
Once you understand the syntax and flow of SELinux policy then writing it is easy. Writing GOOD policy on the other hand …. Lmao.
Typically most IT departments “fix” it with
setenforce 0which is the equivalent of removing the seatbelt cuz you can’t figure out how to latch it.Android has one of the most “robust” applications of it but it doesn’t serve the purpose a good policy does, it does add a substantial layer of defense. Apple contracted my company to come out and teach them how to SELinux a few years back. Ultimately they (companies that desire SELinux as an added layer of defense) tend to just pay “us” to do it instead lmao.
It's open source - literally the opposite of black box.
The "where I live" part is key. Because very likely this person is in SF, where they cannot buy a luxurious house cash with that money, and where cost of living eats surprisingly far into that stupid high number.
But notably, this is why all the normal people who don't make a half million dollars a year can't live in SF! 😅
That's why I mentioned "where I live". 😉
Big tech pays large amounts of money. This is why people choose to work there.
Well yes, that goes without saying...
PS: 🇸🇪👋
Career levels at Amazon (basically pay grades)
Total Compensation
At Amazon you have the following levels
L4 - Junior. A new grad. Expected to be promoted within 2 years or let go
L5 - Mid engineer. Very wide band. Encapsulates anything between a level 2 engineer and a team lead at other companies. Can be expected to lead individual teams at times. Is considered a “terminal” position (there’s no expectation of a promotion past here)
L6 - Senior. Has the scope of what a Staff engineer would at other companies where you’re not only concerned with your team but others in the department. I think like 10% of engineers ever hit L6
L7 - Principal Engineer. You have like 1-2 of these per department. These are more like architects at other companies. About 1-2% of engineers ever hit this band.
L8 and beyond are for fancy hires and shit. Very few if anyone ever works their way up to those bands.
So, where are L1-L3?
Are L3 student programmers?
L2 people who never coded anything in their life?
L1 are people who can't read? Like babies?
Non-engineering roles I think.
IIRC levels correspond to all employees across the company.
Yeah it’s weird and I don’t get it either.
Non-engineering positions
I fought for getting a 4/5 rating at an old job and gave lots of examples. Their argument was that I didn't deserve it because those were just expected. I pointed out my work compared to others in my team and was told that it compares across the company, not the team. I kept causing a fuss about it because I was so angry about it and finally my manager said something about the bonuses has already been communicated and people would be angry to get less. I was confused because I didn't want more money, I was just offended they said I was performing on average when I was going above and beyond every day. It was also really embarrassing to me. If they'd just said the rating doesn't affect anything except your bonus I wouldn't have even cared.
The whole thing is all BS.
That's because they have a fixed budget and the proportions are tied to evaluated performance tiers, increasing your rating would contractually require them to compensate you more from the same pool of money
You're falling for the "we've constructed this machine to tell you no so you can't argue with us" ploy
No, as I said to another, upper management has every opportunity to fix the budget. Your direct manager however can not
I've found that laughing in their faces and putting in 2 weeks is fairly effective at breaking that wall. Amazing what money they can find when faced with the alternative. Otherwise, the correct move is to actually leave. All of you cowards that submit to the machine make it worse for everyone.
And I have in fact left that kind of jobs myself. Not trivial in a job market like this one though.
Need to make unions stronger again.
Nah, that's bogus. It's a private company, they can do what they want. They could have absolutely given OP the 5/5 rating, and just had them sign something saying that they were content with the bonus appropriate to a 4/5 rating. No one would have had to receive a penny less.
It's very annoying to have managers say their hands are tied when they very well could go to bat for you with their superiors. I was lucky to have one manager really push for me in the past like that. It's rare.
That fixed budget is what they always say. The budget for the company is their problem, not yours.
Upper management can certainly increase the budget. Your line manager probably can't
Yeah exactly. So it keeps everyone in their place, powerless to change anything within the structure that exist. Exactly as intended.
Yeah, no shit, thank you for repeating what I said. The point being I never cared about the money and didn't even understand it was only about the money. I only wanted recognition.
At one job, my manager had a spreadsheet that he was tapping away at during my review. He had the audacity to tell me that he had to downgrade some things so that he wouldn't have to go to a committee to defend at the individual or group level.
I transferred to a different product.
You will soon become just as jaded as the rest of us, and stop expecting your company to appreciate you. It wont feel good but you can change jobs often and get your salary up without any feelings of illusional loyalty.
Haha, the same. Was doing great, supported customer calls, onboarding new engineers, along with ongoing incoming tickets and got 3/5, wrote a few good and a dozen bad RFCs.
Then the manager had the audacity to ask why I am changing the company with a 40% raise. I could've asked for promotion, he said.
I don't work at Amazon, but we have a similar system. I've gone all-in on a couple of subordinates saying they deserved a 4/5 for this or that work. And because they were new-hires, I eventually got the grades punched through after a bunch of hemming and hawing.
Also advocated for my own higher-than-average marks on a few occasions. And just arguing the case gave me the grade as often as not. If everyone in the department had been as stubborn and insistent, I don't know that they'd have given the whole floor these grades. But the squeaky wheel...
Yeah this was my experience when I worked there. Driving goals and doing good work isn’t enough. You need a fancy project to demonstrate “expanded scope” otherwise your promo would get rejected.
Sometimes things worked the way you wanted and people got promoted doing their normal job. A lot of times though there were a lot of fancy projects built to get people promos that suckers got stuck with the bill on.
This ain’t a case of one dude scamming the system as much as it is institutional rot from red tape.
Its pretty well known that "lines of code" is a horrible metric to judge programmers with. It seems "number of new projects" is pretty similar, though at a higher level of abstraction.
Unfortunately that metric is applied to a lot more than just programmers; and I think getting rid of it would involve completely restructuring the type of activity our society is oriented around, and would run up against the life philosophy of the people in charge.
Of course I'm not against progress, but I'm talking about executives that don't plan beyond the next quarter, politicians that don't plan beyond the next election cycle, the endless pursuit of growth, and the inability of market economies to cope with the fact that sometimes inaction is more advantagous than action. All of this encourages endlessly churning out 'new' things, without designing those things to last or putting in the effort to maintain them.
Is that why they are gradually replacing the bad AWS Console UI with something 10x worse?
Apart from being slow, having discoverability issues, not being able combine filters and actions so that you frequently need to fall back to shell scripts for basic functionality, it being a complete PITA to compare things between accounts / regions, advanced functionality requiring you to directly edit JSON files, things randomly failing and the error message being carefully hidden away, the poor audit trail functionality to see who-changed-what, and the fact that putting anything complex together means spinning so many plates that Terraform'ing all your infrastructure looks like the easy way; I'll have you know there's nothing wrong with the AWS Console UI.
Funnily enough I joke all the time that the new UI is a subtle ploy to get us to terraform everything. Still PITA when you want see what is fucked up in an API Gateway route.... So... Much... Wasted... Space
Probably, it's also why Google has 100 different chat clients.
Perverse incentives combined with underskilled management 😐
You get the behavior your incentives encourage, whether you realize what those behaviors are or not.
[ In lieu of a comment, please see "Bullshit Jobs", by David Graeber, which is incorporated here by reference. ]
Something I find cool about this book is that it's so well known that people who haven't even read it will often gesture towards it to make a point. It reminds me of how "enshittification" caught on because so many people were glad to have a word for what they'd been experiencing.
It's a useful phrase to have. Recently a friend was lamenting that they'd had a string of bad jobs, and they were struggling to articulate what it was that they wanted from a job. They were at risk of blaming themselves for the fact that they'd struggled to find anything that wasn't soul sucking, because they were beginning to doubt whether finding a fulfilling job was even possible.
They were grasping at straws trying to explain what would make them feel fulfilled, and I cut in to say "all of this is basically just saying you don't care what job you have, as long as it's a non-bullshit job". They pondered it for a moment before emphatically agreeing with me. It was entertaining to see their entire demeanour change so quickly: from being demoralised and shrinking to being defiant and righteously angry at the fucked up world that turns good jobs into bullshit. Having vocabulary to describe your experiences can be pretty magical sometimes
IMO if your survival depends on doing a 'job' (especially if you're employed by someone else), then it's better to look for fulfilment in your personal life and realize the job is a means to survive and hopefully also fund what you really want to do for yourself and your loved ones.
Work to live, not live to work.
Indeed, that is the healthier way to go about things.
Personally, I struggle with that kind of compartmentalisation, but I would probably be healthier if I could do that. I have never lasted long when doing work that I'm not passionate about, and when I am passionate about work, it's hard to not bring it home (even if that's just working on stuff adjacent to the task).
I know a lot of people who work in academia, and it's simultaneously inspiring and depressing to see how people's research interests end up bleeding into basically all elements of their regular life. I think some people are just wired that way. I wish that they had the freedom to engage in that in a more healthy way, free from the additional bullshit that Capitalism heaps onto them, making the dynamic so toxic.
However, given that we do live under such oppressive economic conditions, "work to live, not live to work" is an essential mantra to aspire towards, especially the people who put their whole heart into their work. It's not ideal, but it is necessary to learn if we want to survive without burning out.
Yep, this is the culture I keep running head first into as I try to level up my career.
Same. Generally speaking our company is pretty healthy, but we're still stuck in this really stupid leveling system where advancement is tied to greenfield development and I've been doing maintenance and compliance work for the last five years.
Man, it's so frustrating. I just can't turn off my own qualms with shitty corporate culture and it means I will be less successful by the metrics we've set ourselves.
It’s ironic how so many of us find ourselves being extremely valuable for the exact reasons they can’t stand us. As IT, I’m used to being seen as nothing more than red marks on a budget to the folks making decisions. The only thing they hate more than listening to us, is when they have to.
Kinda got a chip on my shoulder today it seems.
The whole thing is pegging my BS meter, including letting an L5 deploy without a code and architecture review, TC, and the fact that they're posting this and claiming they're still there.
I've got a few friends who work at Amazon, and while the story certainly sounds embellished and a bit too "just-so", the corporate attitude of make-work to justify a promotion even when its a waste of time and resources rings true as a bell.
Did this guy actually oversee a fully transition to a new service and waste a bunch of internal time and money for a system that's sub-optimal by any conceivable measure? Idk, maybe. If he'd just written "Twitter" instead of "Amazon", I'd have taken it at face value no problem.
Did this guy author an overly-complex plan as part of his promotional material, get it vetted and reviewed and rubber stamped by a bunch of friendly higher-ups because they wanted to justify his promotion, and then stuck on a shelf marked "Maybe we'll do this in 2029 if we're not busy with something else"? Equally likely.
Does Amazon have a bunch of bread and butter break-fix work they could be dedicating staff to, rather than chasing the next digital White Whale so they can feel cutting edge? Yeah, no shit. Absolutely.
I believe it. I don't work at Amazon, but I've seen proudly launched pieces of shiny crap support promotions at other companies.
I've seen some garbage slide through code reviews. Most people don't do them well.
I'm doing contract work at a big multinational company, and I saw a syntax error slide through code review the other day. Just, like, too many parenthesis, the function literally wouldn't work. (No, they don't have automated unit tests or CI/CD. Yes, that's insane. No, I don't have any power to fix that, but I am trying anyway). It's not hard to imagine something more subtle like a memory leak getting through.
In my experience, people don't want to say "I think this is all a bad idea" if you have a large code review. A couple years ago, a guy went off and wrote a whole DSL for a task. Technically, it's pretty impressive. It was, however, in my opinion, wholly unnecessary for the task at hand. I objected to this and suggested we stick with the serviceable, supported, and interoperable approach we had. The team decided to just move forward with his solution, because he'd spent time on it and it was ready to go. So I can definitely see a bunch of people not wanting to make waves and just signing off on something big.
I can't speak to this situation, but broadly speaking I am familiar with general messed up stuff like this as well as perhaps adjusting some fine details to make the scenario relatable to an audience unfamiliar with the specifics of the real situation and/or obfuscating the details so that the person doesn't out themselves to someone else familiar with the specifics enough to recognize.
The broad strokes seem plausible and any oddities in details I consider to be less important and/or understandable if it was tweaked for an internet audience.
This almost makes me appreciate my current job, where most stuff has been in place for years and any changes take forever.
It's kind of a bummer that it's going to take like six months to add a linter, and they only started using git like last year.
I worked in a heavily regulated industry. Everything required a manual test. Let's say you have an employee ID that is 10 digits long which they use to log in. You had to have some else (couldn't be the developer) to write a series of tests, get those tests approved by 5 people(with specific titles) then a third person to execute the test, then the second person had to write a report saying it all passed, then that report had to be approved by the same 5 people.
That typically wasn't the delay. The delay was to execute the tests we needed to stop production. That typically was a 6 week wait(unless urgent for "reasons") and changes like "I will drop scrap by 83%" was typically told wait till July 4th or Christmas breaks. Why? Because production would be down for 3-4 days typically. Someone had to start the system, ok no entry produces error, executor and developer have to sign a physical paper, restart the whole system, now an entry of 1 digit produces an error, sign the form, repeat for all digit quantities up to 9, repeat for all digit quantities up to the choosen value(based on severity if an issue occurred), 2 people sign for each one, system restarted between each. If you had say an enter button and a cancel button each had to be checked for each quantities of digits. Oh but wait what if someone just types there name... Now repeat everything for alphabet values... What if someone does combination, more tests, more restarts, more signing.
Reports easily surpassed 1000 pages, no one really had time to check all that so I saw so many missed signatures and missed tests. I asked the "senior validation expert" can I just automate a lot of these tests using unit tests and attach a computer generated report of all tests passing and the source code of the tests? " the response I got was" what's a unit test? "they still don't use any of them to my knowledge.
Similar boat, it's kinda frustrating that it takes 6 months to approve 30 minutes of work, but at least the job is boring.
Which VCS did they use before git?
SCP to prod, or ssh in and copy paste. Devops only removed write access to prod machines this month, and people complained. (No, we don't have docker)
I think they used Amazon CodeCommit for a while, but I don't know what that's like.
Wow, I imagine that caused a lot of problems.
You would be surprised how far that type of thing can get you when the team is small and experienced.
It tends to explode when you hit a certain number of people or you replace a senior with a junior who promptly explodes the thing.
we dealt with similar stuff at our company as the design team grew. amazing how far simple systems can get you with basic practices and common sense.
now with triple the team size and a few less than extremely competent people, we have tons of file management issues, even though there are more processes in place to avoid them. I hate it.
I'd ask if you're at my work, but this is an amount of organizational improvement that they haven't yet been able to begin
Sounds about right. There is no longer any incentive to focus on maintenance and incremental improvement (the stuff that actually keeps the lights on and the revenue flowing). It's all about the new and shiny--even when it results in regression.
Which is why AI and vibe coding will survive. Besides the part where it’s not my code, the company owns it. The fuck do I care how good it is. If it works and gets me a promoted or moved to a new spot in a different company. Heck yeah. Issues down the road are not my problem.
Personal project: page load takes under 10kB and any button or link loads in millisecond.
Work project: Fuck it 26MB page load. It's not like the pages load in under 5s before anyway.
The personal project is a matter of personal pride, whereas for work, any old thing will do, as long as it meets the requirements.
I apologize for bashing Java so hard in the past. I wish everyone wrote everything in Java these days. Digital life would be so much better.
You owe royalties
Yeah, Oracle licencing has really taken the shine off Java and relegated it to the legacy dust bin.
Fuck no.
I wish everyone used C#, Scala, Rust or Python (DSLs like VHDL, SQLs and CUDA and super specific languages like C, Erlang, Haskell and Bash notwithstanding).
You can hate on them, sure, each for their own reason, but they're all very well supported and good for what they're intended for.
Sure, me too, but that's my point. Even Java is better than what we have now, especially from the user's perspective.
I say this is only ok because he did that in amazon. Fuck amazon
If he did that in a medium-or-less sized company that would be a really shitty move.
In a small company noone would try to label you "l5" or "l6" and probably an actual human would make your comp decision. You take the byzantine incentive structure away and people just try to do a good job.
The problem is the large companies like Amazon buy all the small ones and put these people in charge lol
Ok, now fire him.
The fact they had to do this to earn a promotion is an institutional problem. Don’t hate the player, hate the game.
I can do both, tbh. Though I do generally hate the game more than the player.
Game rules: You want a promotion? Make something cool, improve something while using approaches that will show that you deserve a higher position and, therefore, a bigger salary.
Player: (Lies and creates shit that is even worse than the initial situation.)
Lemmy: Don’t hate the player, hate the game.
You are contradicting yourself. If writing bullshit and making things worse gets you a better career position
Is not the rule of the game. Sell your story to your superiors is the rule of the game, that's the real metric, the the thing that really matters.
The scumbag behavior is from the employer. He's only fighting fire with fire.
Do you want me to present you with a definition of "lie"? I believe you don't understand the phrase "Lies and creates shit".
They built something worse and we're still promoted for it despite it being demonstrably worse. Where's the lie? They described something complex and techy sounding, did it, and got the promotion anyway regardless of the actual results, proving the results didn't matter.
So you want the manager to be cleverer than the engineer in engineering, so the manager would be able to detect a deliberate lie from the engineer?
Yes, but more competent, not cleverer. Some managers aren't fit to be in IT.
I'd expect a manager to be able to determine that testing data for the new process is showing it is worse than the previous system it replaced, and NOT promote that person, at the very least ...
More like game rules: manager needs shiny buzzwords and big number go up. Having something that works fine for 5 years is considered stale and corporate culture is all about useless innovation.
But that isn't the game rule, now is it?
The rule is more: convince the c-suite that you deserve a promotion by any means necessary. Even if you have to make things up.
This is the difference between RAW and RAI.
the player can always choose not to play, though
It's a big company. Someone's going to play.
No. That putting the onus of change on individuals is a losing proposition. The incentives have to change or no number of good people will fix it. I hear the French have had very effective solutions in the past.
You seem to believe that I think it a justification for evil. I do not, people should not do such things and they are shitty people for doing them.
I'm saying that the idea of some good people doing the right thing fixing the problem is naive and doomed to failure and a real solution to the problem has to be bigger than the lazy "just no one be evil" proposition you seem into to champion.
Fire the manager too.
The one thing which COULD justify it, is technical debt. A programming language not supported anymore or in short-term/mid-term, bus factor, too much knowledge transfer, etc. But yeah, lots of times it's "business as usual" just for "progress" and fancy buzzwords.
Java is still supported... Or did I miss the memo?
What's Java???
Golang is technical debt in language form. A language that gained limited and now sagging popularity, for good reason. I hate to work in Java but hate golang more. It's the lightsaber of programming languages. I've got shit to do, give me blasters and all the rest. And I'm not interested in wanking myself off about how I did it all with channels. [edited for typo/clarity]
Flawless victory.
All of you who can't understand the concept of feature complete (see syncthing drama for example) find something useful to do. I promise it's out there.
This is directly caused by squeezing promos down to making your skip level manager horny for the sound of your work instead of having any actual impact on what you are doing.
Its all such stupid horse shit.
"Why does X feature in Y app/game/device take so long??"
This is why
I am so tired of worse products in the name of upgraded products that are literally worse in every way but a bunch of buzzwords and in groups bragging at the top while not knowing anything at all about programs or even the product at all but just seem to be there because they drink with the CTO.
Ugh. The twiddling thumb era of trying to look busy by dismantling the old machines for parts.
So that's why we suffer enshitification.
Those who succumb to the Socio-Economic and climb it so.
"Upwards mobility".
Honestly probably got the project to more maintainable state. Probably didn't need the rewrite to do it in a new lang to do it (the real killer hear it sounds like).
Those monoliths suck on the operations side, and even worse when it's a corpse holding up the foundation to other projects that actually need it to change. Need to scale? good luck that decades old pizza box we call a server isn't supported anymore. Oh of course we can spend millions virtualizing dead hardware to keep it running the same.
Yeah, longterm wise - Go is far easier than Java to maintain. This is still a win, albeit with a slight initial disadvantage
somebody please hack into amazon's services so that they can tell amazon shoppers the truth about jeff bezos. seriously!
https://youtu.be/gwicyfJbYD4
Swap Java and Go in text, then I buy it. Java is memory hungry monstrosity that runs on JVM and idiomatically uses piles of abstractions. I have exactly opposite experience, when rewriting a microservice from Java to Go reduced memory usage tenfold and sped up requests processing.
No, I believe it this way. It used to be one service that had access to everything it needs. Now it is microservices, so each microservice is caching a bunch of stuff, but of course all the wrong stuff, so every request requires at least one network call downstream. Thus more memory usage and slower.
Well, that's the architecture problem, not the language.
With a little effort, one can write bad code in any language.
According to votes, hating Java is bad, but hating microservices is good.
Java's biggest strength is that "the worst it can be" is not all that bad, and refactoring tools are quite powerful. Yes, it's wordy and long-winded. Fine, I'd rather work with that than other people's Bash scripts, say. And just because a lot of Java developers have no concept of what memory allocation means, and are happy to pull in hundreds of megabytes of dependencies to do something trivial, then allocate fucking shitloads of RAM for no reason doesn't mean that you have to.
There is a difference in microservices between those set up by a sane architect:
... and the CV-driven development kind by people who want to be able to 'tick the boxes' for their next career move:
We mostly do the second kind at my work; a nice Java monolith is bliss to work on in comparison. I can see why others would have bad things to say about them too.
I was referring another comment in the thread, sorry for confusion. The OP attacks both Go and microservices, although it's no Gos fault in the story.
Also I just hate Java too, and OOP in general.
If the Java monolith used 6 gb of ram and each Go service uses 500mb of ram, but now there are 16 micro services, which uses more memory?
Monoliths are more efficient, but hav other issues.
Can be more efficient, but it's not always. You can scale those services more too. Do you need all 16 running on idle? Lazy load then as needed. Also does it really need a whole 500mb? Why? If its just a cache made that's better handles with another shared service (redis, etc). If it's software, why? A full "fat" Suse Base Container image uncompressed is 94.8mb and not all of that will be loaded in RAM. Going down to a micro at 23mb. All this and now you can deploy it on the shared infra, no separate OS, Management/security/logging/networking solution.
Sure, but they also have duplication, IE database connections. Each service needs at least 1 connection versus one for a monolith.
There are pros and cons to any architecture.
The IPC increase is the most sure tradeoff for sure.
Agreed, engineering is deciding what trade offs are acceptable. Cargo culting any pattern is a good way to build runaway junk