The ledger.

Where things stand

Some of this is uncomfortable to write down. It is still better than you finding out on your own.

• An item loss bug on Elwynn

  waiting on a window

  Characters were losing items that had been in their bags. The cause is understood and the correction is already in the configuration, but the running server has never read it. It takes a restart to take effect, and that restart has not been scheduled yet. Barrens is not known to be affected, and we are careful with the wording: no confirmed case in the data we checked is not the same as cleared.

• Double experience for brackets already opened

  waiting on a window

  Not built yet, and it does not belong to this gate. Catch-up runs in the run up to the last phase of an expansion and in the pre-patch into the next one, because those are the points where everybody else is at the cap and starting late actually leaves you alone. Between two levelling phases the brackets behind you are still full.

• Nobody is woken up when the realm goes down at night

  open

  There is monitoring work in progress, and until it lands, an outage outside our waking hours is found by players before it is found by us. If the realm is down and quiet, telling us on Discord is genuinely the fastest route.

• The gear planner shows items that do not exist here

  open

  A player reported it and the player is right. The planner was built from a snapshot of the other realm, taken before Barrens existed, and it does not know which content stage this realm runs on. So it will happily offer a level 19 character raid gear from a patch this world has not reached. It is on the list, and the fix has to be tied to the phase schedule, otherwise it goes stale again the moment the world moves on.

• The fishing quests a player could not find

  done

  Also correct, and also not a bug. Those quests belong to a later content stage than the realm is standing on. They arrive when the world reaches that stage.

Five things we had wrong, caught before they shipped.

Every piece of work here gets read back by a second, independent reviewer before it goes anywhere near the realm. These five were caught that way. They are listed because they share one property that is worth understanding.

• We built a health check that could not see the crash that prompted it

  done

  Knocking on a service from the outside tells you nothing when that service restarts itself in seconds. The check would have been green through the entire incident it was written for.

• We planned an approval that would have approved nothing

  done

  Two different things were confused: who is allowed to log in, and which connections are allowed through. Fixing the first would have left the second exactly as shut as before.

• We renewed an alerting account and forgot to write it down anywhere

  done

  Every alert would have gone silent, while the setup itself reported success.

• We wrote a guide that would have quietly disabled the feature it enabled

  done

  It replaced a configuration file in a way the running service never re-reads. The new value on the outside, the old behaviour on the inside, and no error message anywhere.

• We claimed Barrens was unaffected without being able to show it

  done

  The chat exports we were reading end before the realm opened, so the post launch discussion was simply missing from the material. The honest version is no confirmed case in the data we checked, and that is what we say now.

Every one of those five would have reported success. No error, no warning, nothing red. That is the failure mode worth guarding against on a server that people trust with their time, and it is the reason nothing here ships because it looked fine.