Yep and it's a job. If it was me I'd tell them I have be paid an hourly wage as a contractor to come do that kind of thing. The pitiful thing here is not that they lost coprorate historical knowedge, but that they never documented so much in the first place. Sadly this is not something that I can blame on DayPaw since this shortcoming is endemic in the game industry and is not uncommon in commercial software in general.
Well I program for a living and the first thing you forgo completing on a deadline or do after the completion of the project is documentation. You really should be commenting/noting your work while doing it so it is fresh in your mind. I see it to often and it really does cause issues down the line. In my case I have gotten so old my memory is no where near what it use to be and have to decipher even my own. Anyways looks like you all have some growing pains ahead of you, good luck.
Thanks for confirming what previous developers have said about causes of lag over the last couple of decades while they kept adding more swarm pets, weapon procs and increasing melee hits / second. Delete 90% of swarm pets and spell/weapon procs, nerf haste / multi-attack and then keep increasing the low-cost player stats like attack, base-dmg, damage bonuses etc until classes are balanced and players can defeat raids again. Don't be afraid, it'll work out fine!
I like this post, and the transparency. Step 1: Adding middleware to send telemetric data somewhere for analysis (New Relic is super awesome for this), is a positive step. Step 2: Aggregating log data as well, in some sort of seachable (lucene-based) engine, like Sumo Logic, Splunk, etc... Those are always the first steps, right? Knowing when and where something is sick, and being able to drill down to see why. Based on what you described, it sounds like three other steps now could even help. Step 3: Add a distributed Cache? Or even a two tier cache for storing real-estate data? If you could leverage something, like Redis as your distributed cache cluster, I would not be shocked if the 60 ms times drop to single digit nanosecond times. I am not sure you can use a near-cache (or if that would be exploitable), since I do not make games... Wish i did, but that is another story... But if you can store data scoped even briefly in near cache... Step 4: More asynchronous programming? And handling cancellation tokens, for when things timeout after 500 ms, the entire collection is aware and stops processing. Step 5: Maybe doubling down on adding new unit and integration tests in troublesome areas of the system. I know the POs would scream at you, but if it is taking a week to release a fix, maybe the investment is justified... I know, i am on the outside looking in, i know nothing. But... for what it's worth... as a fellow software engineer with 27 years background my two coppers.
Considering Bristlebane server players have had several months of it taking 30-40m to move a raid from Zone A to Zone B with cascading crashes every single time, yes - Bristlebane players have had marked improvement when it comes to our World Lag since the patch. 30s is much better than 30m (10m+ per zone plus crashes). Whether this improvement continues through the month remains to be seen. Zone lag is still problematic for all.
Earnest communication is good, no matter how you slice it. Of course this invites every Tom, , and Harry to comment on how they would fix it. Thanks for taking the time to post. It's HUGELY appreciated.
Raids have become entirely unplayable for me due to spell lag and not able to cast spells. This is new and only been happening from about the time of the 64 bit update. Its super frustrating. I hope its fixed soon as there is no reason for me to keep trying to play the way it is.
All swarm pets are not only dps, they are also tanks. Making them dots would only add to the buff cap problem on raid mobs. For swarm pets that are not tanks i would be fine with removing them. They can be replaced with a DD AA or whatever triggerd them. Pretty boring but if it helps im fine with that. For swarm pets that are tanks i would spawn just one and make them as powerful as the number of pets that are currently spawned. For procs in general i would reduce the proc rate dramatically and increase the damage.
This one may be unpopular as well, but.. perhaps treating static zones similarly to dynamic ones? Plane of Knowledge may see constant use by hundreds of folks at tall times, but original Vex Thal is empty every time I go. I'd happily take waiting for a low-demand zone like VT to spin up the way I do a seldom-used guild hall if it mean the server didn't need to be completely aware at all times about the 700 mobs inside it. Changing the classic guild hall into a static zone everyone can access might be worthwhile too; standard bankers already only show you your personal content, so it doesn't seem impossible to make guild bankers do the same.
If you are suggesting this just for access to guild bankers wouldn't it be easier to add them to an existing zone such as the guild lobby?
There already are bankers in the guild lobby. And the standard guild hall off of the lobby is part of a quest and used by the in game guild. Guild bankers have the additional filter to allow certain things be public or just for guild members to purchase etc I think that would mess with too much. The only way you can move the masses to a different under used zone is to add something everyone wants. So add buff bots to some out of the way zone and players will flock to it. Even if you just offer buffs for those 60 and under or what ever the cut off is. Then put some port features there and you have a popular zone. Lobby #2 heh
We know the symptoms. A post listing the symptoms just lets us know you know them, and at this point it'd be scary if you didn't. Then a list saying processing data makes loads on systems is like no kidding? So, solutions? Don't see any real talk here; just history.
Bag sales make money (but expand the request increasing lag.) Dragons hoard is small unless you buy more slots. I feels like the answer to your question is that they decided it was better to monetize it than to make it free and address the problem it was implementing to address. Kudos to the devs. Shame on marketing and management???
Dragon's Hoard is still small even when you do buy more slots. I can have 3 houses with 300 items each (900 items) in my houses - of any type; in my case, tradeskill items. A dragon hoard with all slots unlocked is only 225 items. Paid or free, people aren't going to give up their houses for storing things until Dragon's Hoard can store 900+ items.
