Miller didn't crash tonight, here's why.

it could be software related. if you look at tthe graph it never crashes below 2000 but not always at the same number. probably something with the queue when Indar is already at 2000 players (of course that is only valid if they raised the pop cap per cont back up to 2000)

Ah so this is why Sony is chasing smart players away by making all their creative playstiles useless?

(quoting you just so that it flags as a reply and you see this)
That graph you posted, are you able to do another for the last 7-8 days? or between certain dates of the 10th through to the 16th?
Given it crashed nearly every day last week it would be good to see the week graph to see if it shows the same trend of crashing at 2000 players just to further back up the claim.

If it was software side why the mapping to actual hardware. 4 other servers on at least 3 other physical machines are having no problems (this is based on the assumption that Matherson and Waterson 'could' be on the same machine, Connery and Briggs could not).

The problem is entirely local to that one server in amsterdam and its conking out relative to a certain amount of load. Thats far more likely a hardware error.

The indar que is an intesting theory, we know things like being in the tutorial have brought the PTS down before (though it was pointed to this being due to it having weaker hardware so..) but if thats the case why is it not as prevalent on other servers like ceres which presumably (due to the indar effect - indar fills up first) are having just as much stress on their indar que.

Also if it was software not hardware, its strange that it presents itself 2 days after a patch, not after the first 'load' situation after the patch.

It could be any number of things, but if its hardware it makes sense given the geographic and (likely) machine locality, if it was software based, Connery is hitting numbers far in excess of any of the EU servers.

I haven't even wanted to try logging in for the last week. The server will either be down or doing the whole 'logging in' for 20 minutes thing... By which time the server will have crashed. Ridiculous.

Would say its about time SOE start adressing that issue as well

i already thought of the EU/US situation, i just had to retype because my crappy phone crashed.
one dev said on twitter that there was a memory leak from the EU servers to the database (or whatever it was) in the US. it could be latency related

i'm not really familiar with the subject, those are just some guesses based on the graph and some logic

Disagreements aside, this needs to stay on the front page and near the top.

No ones in the position to rule out anything, its simply orders of magnitude more likely to be hardware based on the available evidence.

Here, note the first "dip" is actually the downtime for the valentines day patch. Since then, Miller has crashed everyday since yesterday... and so far tonight:-

Actually, the 2000 mark means nothing. Since last thursday patch, it has crashed several times with sub 2000 pop (mostly around 1800 server pop total). It just been the last three days where it has crashed at the 2000 mark.

If i look at the usual population it looks like the only time it crashed with less than 2000 players online was 15. the first crash, where it should have been above 2000. especially as the growth would suggest that there should have been more than 2000players at that time, probably due to the increased login time.
i can only guess that the first 2 crashes occured before they took down the leader board, which would kinda lead to the conclusion that there is infact a bandwith problem when more than 2000 players are online (or equal load with the leaderboards online)