Server settings and maintenance
Platform settings live in the global admin portal under Settings. A few server-level options are set in the server's configuration file instead.
Settings in the portal
| Section | Holds |
|---|---|
| General & branding | Brand name, accent colour, default theme, server name, default region and timezone |
| Security & access | Enforce MFA, session timeout, minimum password length, single sign-on |
| Team & roles | Your staff and the permission matrix. See Users and roles. |
| Notifications | Storage warning level and the Slack, webhook and PagerDuty channels. See Monitoring, alerts and reports. |
| Backup defaults | Default retention for newly enrolled devices |
| Storage & retention | Replication to secondary targets and the egress guard |
| Billing & licensing | Currency and per-seat and per-GB rates. See Billing tenants. |
| Maintenance | Operating modes, key rotation and factory reset |
Changes to settings take effect within a few seconds, without a restart.
Maintenance modes
- Maintenance mode pauses backups and restores: agents' live connections and the restore and download endpoints answer "unavailable" until you turn it off. Use it while you work on storage.
- Read-only mode blocks configuration changes in the global admin portal, apart from turning read-only mode off again. Use it to freeze the setup during an audit or a migration.
Rotate keys
Rotate keys asks every agent to start a new encryption key on its next backup. See Encryption and key recovery for what it does and the full re-upload it causes.
Factory reset
Factory reset erases every tenant, storage target and per-tenant setting on the server. To confirm, you must type the server name exactly. It does not delete objects already in storage.
Export the configuration
Settings > Maintenance can export the platform settings as JSON, for your records or to compare two servers. Secrets are not included.
Back up the server itself
The server's database holds tenants, users, devices, restore-point records and settings. Without it, agents keep their data in storage but nobody can sign in or find restore points.
The server can dump its database on a schedule, keep the newest dumps (seven by default) and upload each one to an S3 bucket. Turn this on in the server configuration and point it at storage off the server. Keep a copy of the server's configuration file somewhere safe as well. Storage credentials and other secrets in the database are encrypted with a key from that file, so a restored database needs it.
Health checks
| Endpoint | Use |
|---|---|
/healthz |
Liveness: the API process is up |
/readyz |
Readiness: the API can reach its database |
/metrics |
Prometheus metrics, protected by a bearer token you set in the server configuration |
Audit log retention
Audit events are kept forever by default. To limit them, set a retention period in days in the server configuration. Older events are deleted once a day.