Reliability
The public entry point for Panda incident and availability updates.
All production components are expected to be operational unless an incident is posted here or announced through support.
Panda monitors Discord gateway health, HTTP readiness, SQLite readiness, queue depth, SOL payment verification, quota checks, managed AI availability, web search availability, and music sidecars.
Core availability depends on Discord gateway events, Discord interactions, the landing and billing pages, the assistant API, queue workers, billing verification, search providers, managed AI providers, and storage.
Music availability may also depend on sidecar health, stream extraction, network conditions, and Discord voice connection state.
During an incident, Panda may degrade paid AI responses, web search, schedules, or music while leaving help, billing, export, delete, and support commands available.
Incident updates will avoid raw server content, provider model names, API keys, billing secrets, and private customer details.
When an upstream provider is unavailable, Panda may return a degraded response, queue work for later, pause spend-heavy features, or ask admins to retry after service health improves.
Billing and entitlement checks remain conservative during degraded states so server owners can see plan state before paid work resumes.
