Sign in
settings

Update Maintenance Mode

Freeze or unfreeze dashboard sign-ins, for this and every other replica.

The new state is persisted and nothing is applied to the running worker, because every reader goes back to the stored row. That is what makes one call enough for a deployment running more than one of them.

patch/v1/settings/maintenance-mode

Request body

enabledboolean required

True to freeze new dashboard sign-ins, false to allow them again.

Response

Successful Response

enabledboolean required

When true, POST /v1/auth/session refuses every credential with 503 so nobody starts a new dashboard session during a redeploy. Sessions already issued keep working, and the management API and the data plane are unaffected: a caller presenting the master key or an API key through the header is never frozen out.

Changes