Operational work changes shared state. Reading code, running static checks, and building an image are ordinary development; migrating a database, deploying a release, changing runtime configuration, or moderating real content requires explicit authority for the named environment.

New volunteers can prepare, review, and verify the non-production parts of an operational task. This handbook does not currently enable independent production operation: it cannot supply the missing provider access, environment map, backup procedure, or canonical rollback command.

The repositories describe much of the mechanism, but they do not currently contain a complete production deployment, credential, backup, or rollback runbook. A release or incident task must therefore record the missing environment-specific facts before action starts. Do not fill the gap by guessing from an old provider name or a local .env file.

Every operational task names

  • the service and environment;
  • the current deployed revision when known;
  • the person carrying the next action;
  • the exact permission being used;
  • the expected user-visible effect;
  • the verification method;
  • the stop or rollback condition; and
  • where the result and follow-up will be recorded.

Production access does not make an action routine. Prefer a reversible, observable step; keep unrelated changes out; and leave the system easier for the next volunteer to understand.

Housekeeping is not maintenance

Packbase maintenance mode is a product-operational state exposed by the server. Wild HQ Housekeeping Mode is a governance pause for a foundational disagreement. An outage does not automatically trigger Housekeeping Mode, and Housekeeping Mode is not a substitute for an incident notice.

→Database and migration changesHow to change Packbase PostgreSQL and Prisma state without treating a destructive sync command as routine development.

→Background workHow Packbase queues, imports, cleanups, retries, and graceful shutdown affect implementation and incident work.

→Releases and deploymentsHow to prepare, authorise, deliver, verify, and recover a Packbase release without guessing its environment.

→Incidents and maintenanceHow to recognise, coordinate, mitigate, verify, and learn from a Packbase service failure.