It’s unfortunately getting worse and worse. The frequency of high-traffic errors and agent terminations seems to be increasing, and the tool is becoming progressively harder to use in any reliable way.
I would also like to add one more practical suggestion that might be worth considering while the root cause of the instability is being addressed.
If bot traffic or automated routing abuse is contributing to the current capacity and rate-limit issues, I would seriously consider temporarily invalidating all long-lived authorization tokens and replacing them with short-lived tokens (for example, expiring every hour).
If necessary, require full re-authentication every 60 minutes until the situation is stabilized.
From a user perspective, I see absolutely no problem with logging in once per hour if that helps eliminate bot abuse and restore stability. I would much rather:
-
work uninterrupted for one full hour,
-
re-authenticate once,
-
and continue working normally,
than have to click the “Retry” button every 10 seconds and constantly deal with popups like:
“Agent terminated due to error…”
“Our servers are experiencing high traffic right now…”
If automated or scripted access is part of what’s saturating capacity, shortening token lifetimes could significantly reduce persistent abuse and make session hijacking or token sharing less effective.
Predictable, uninterrupted access is far more valuable than long session persistence in the current situation.
