Too many idle players

[Nibble]

It shouldn't matter how many people are on.

Idle kicks are not the right answer.

Free world transfers are a partial solution, as are additional worlds. But these too are not the right answer.

Better software and systems engineering is the only real solution.

I have no faith that continued maintenance will solve the problem and I expect this to continue for several more weeks until world transfers are made available. Even then, the problem remains, these services can't handle the load. They bug out.

Yoshida does not guarantee capacity problems will be addressed. They are "hoping" and they "believe" that they can handle the load in a week. The lack of a definitive answer really bothers me, as though Yoshida's staff is unable to define and correct the problems we now experience.

Meanwhile you are clamoring for ridiculous things that will only take time away from making a product work as intended.

I've personally written services that handled millions of concurrent users, billions of requests per day, hundreds of thousands of requests per second on clusters 'small' enough to fit in only a few cabinets (read, a few hundred machines). I can't fathom why Square is still experiencing problems. Is it tooling? Is it platform? Partnerships? Were these services not engineered to scale out on demand? DID THEY NOT HIRE THE BEST OF THE BEST FOR THIS?! Maybe this is another draconian approach to service implementation and topology where you have 100 service instances, all single threaded and all on the same physical network.. e.g. not truly scalable. *shakes head* I've worked for companies that screwed up in capacity planning and watched cluster sizes double three times over, daily, until resolved.. I've watched software utilization drop to 10% as our brightest developers in the company rethought how services should operate and worked tirelessly to solve load and performance issues, I've watched companies overcome with a sense of urgency I'm not seeing come from Square.

...and you think an AFK timer will help solve "the problem" -- *shakes magic 8-ball* "Seems unlikely." During peak hours servers will still fill up for as long as the game is popular, with active players, replacing who would have otherwise been inactive. What then, should we ask for randomized disconnects as well?

</rant>

[Aleisterz]

It shouldn't matter how many people are on.

Idle kicks are not the right answer.

Free world transfers are a partial solution, as are additional worlds. But these too are not the right answer.

Better software and systems engineering is the only real solution.

I have no faith that continued maintenance will solve the problem and I expect this to continue for several more weeks until world transfers are made available. Even then, the problem remains, these services can't handle the load. They bug out.

Yoshida does not guarantee capacity problems will be addressed. They are "hoping" and they "believe" that they can handle the load in a week. The lack of a definitive answer really bothers me, as though Yoshida's staff is unable to define and correct the problems we now experience.

Meanwhile you are clamoring for ridiculous things that will only take time away from making a product work as intended.

I've personally written services that handled millions of concurrent users, billions of requests per day, hundreds of thousands of requests per second on clusters 'small' enough to fit in only a few cabinets (read, a few hundred machines). I can't fathom why Square is still experiencing problems. Is it tooling? Is it platform? Partnerships? Were these services not engineered to scale out on demand? DID THEY NOT HIRE THE BEST OF THE BEST FOR THIS?! Maybe this is another draconian approach to service implementation and topology where you have 100 service instances, all single threaded and all on the same physical network.. e.g. not truly scalable. *shakes head* I've worked for companies that screwed up in capacity planning and watched cluster sizes double three times over, daily, until resolved.. I've watched software utilization drop to 10% as our brightest developers in the company rethought how services should operate and worked tirelessly to solve load and performance issues, I've watched companies overcome with a sense of urgency I'm not seeing come from Square.

...and you think an AFK timer will help solve "the problem" -- *shakes magic 8-ball* "Seems unlikely." During peak hours servers will still fill up for as long as the game is popular, with active players, replacing who would have otherwise been inactive. What then, should we ask for randomized disconnects as well?

</rant>

I would rather have active players then inactive ones just taking up space.

[DayumQuitPlayin]

It shouldn't matter how many people are on.

Idle kicks are not the right answer.

Free world transfers are a partial solution, as are additional worlds. But these too are not the right answer.

Better software and systems engineering is the only real solution.

I have no faith that continued maintenance will solve the problem and I expect this to continue for several more weeks until world transfers are made available. Even then, the problem remains, these services can't handle the load. They bug out.

Yoshida does not guarantee capacity problems will be addressed. They are "hoping" and they "believe" that they can handle the load in a week. The lack of a definitive answer really bothers me, as though Yoshida's staff is unable to define and correct the problems we now experience.

Meanwhile you are clamoring for ridiculous things that will only take time away from making a product work as intended.

I've personally written services that handled millions of concurrent users, billions of requests per day, hundreds of thousands of requests per second on clusters 'small' enough to fit in only a few cabinets (read, a few hundred machines). I can't fathom why Square is still experiencing problems. Is it tooling? Is it platform? Partnerships? Were these services not engineered to scale out on demand? DID THEY NOT HIRE THE BEST OF THE BEST FOR THIS?! Maybe this is another draconian approach to service implementation and topology where you have 100 service instances, all single threaded and all on the same physical network.. e.g. not truly scalable. *shakes head* I've worked for companies that screwed up in capacity planning and watched cluster sizes double three times over, daily, until resolved.. I've watched software utilization drop to 10% as our brightest developers in the company rethought how services should operate and worked tirelessly to solve load and performance issues, I've watched companies overcome with a sense of urgency I'm not seeing come from Square.

...and you think an AFK timer will help solve "the problem" -- *shakes magic 8-ball* "Seems unlikely." During peak hours servers will still fill up for as long as the game is popular, with active players, replacing who would have otherwise been inactive. What then, should we ask for randomized disconnects as well?

</rant>

I Lmao'd @ Randomized disconnects lol.
These ppl are so selfishly anxious that they're wanting ppl who idle above 15 mins to get booted. WTF.
Sorry, not happening. 15 minutes... Ridiculous.