One-time quota reset request after a runaway-session bug (project p_2h4wc79ayet)

Cloud Dashboard
1

Hi LiveKit team,

I’m on the free Build plan (project “Personal-Hermes”, p_2h4wc79ayet).

Between Aug 10-12 a bug in my self-hosted agent leaked five long-lived rooms - agent participants stayed connected after the calls had already ended. The largest session (RM_4RZnyY6gyHo5) ran 2,440 minutes. Together the five account for 6,305 of my 6,582 used WebRTC participant minutes, and the project is now blocked until Sep 1.

The five sessions, if it helps you verify:

  • RM_4RZnyY6gyHo5 - 2,440 min
  • RM_oeqKHeP2Jj5W - 1,199 min
  • RM_wBXmjzzbiraM - 890 min
  • RM_pBaLxTEuEBvT - 889 min
  • RM_TfTB5meVn4kb - 887 min

The five rooms were terminated on Aug 12 when the cap enforcement kicked in, and the project has recorded zero WebRTC minutes since. The code fix shipped Aug 13 (ring/ready timeouts, dead-man watchdog, per-call caps), and as of Aug 15 it’s covered by 14 passing regression tests plus a local circuit breaker that refuses connection attempts while over quota - so a repeat can’t happen.

My normal usage is ~13 minutes/day (~390/month) - far below the 5,000 allowance.

Would you consider a one-time reset or temporary unblock of the August connection-minutes counter? Happy to share logs or any further session details.

Thanks,

Demetre

2

Hi Demetre,

Unfortunately we are not set up for that; it’s one of the benefits of being on our ship tier that you wouldn’t have this disruption.

One ‘workaround’ would be to create a new account, with a new email address, and that is the approach most people take who find themselves in your position.