On this page, you can find how to create, edit, upgrade and delete Databases in the Cloud Console.
The Database-as-a-Service feature in the Cloud Console is currently in beta. It has not been fully tested across all use cases. Customer support is available for any questions or issues you may encounter.
Select Databases in the VIRTUAL DATACENTER block of the side-bar menu. The page lists all Databases of the Project with a search bar, sorting and the CREATE DATABASE button. Click the name of a Database to open its details page.

| Column | Description |
|---|---|
| Status | ACTIVE - ready to use; PROVISIONING - being created; UPGRADING - moving to another version; DEGRADED - running, but not fully healthy (for example, the High Availability standby is not ready); STOPPED - stopped; FAILED - could not reach the requested state, contact customer support; DELETING - being removed |
| Mode | Standalone; Primary - the Database has Replicas in other Projects; Replica |
| Connection string | address and port to connect to the Database |
The Actions menu of a row has Start (only for STOPPED Databases), Edit and Delete.
Click CREATE DATABASE, fill in the form and click CREATE.

| Field | Description |
|---|---|
| Service name | PostgreSQL, TimescaleDB, MySQL, MariaDB, MongoDB or Redis |
| Name | up to 16 characters: Latin letters, digits and hyphens; starts with a letter, does not end with a hyphen. It is also the name of the first database inside the instance, so names reserved by the engine (for example postgres, mysql, admin) are refused |
| Version | only supported versions are offered; the recommended one is marked (default) and is used if nothing is selected |
| Extensions | PostgreSQL and TimescaleDB only: extensions created in the initial database, see Database Management |
| DB user | optional name of the main user: starts with a Latin letter, letters, digits and hyphens, up to 16 characters; names of system accounts (root, postgres, admin, default, replication, …) are reserved |
| Port | the default port of the service is pre-filled |
| vCPUs, Memory | resources of the Database |
| Disk | size in MiB, the range depends on the service; pre-filled with the minimum |
| IP ACL | IPv4 networks with a prefix length that may connect to the public endpoint, one per line, for example 192.168.1.101/32. Later it is changed on the NETWORK TAB, see Database’s Network |
| Provision in private network | places the Database in a private Network and Subnet, optionally with a Private static IP; more endpoints can be added later on the NETWORK TAB |
| Enable High Availability | adds a standby that takes over automatically, see High Availability |
After you click CREATE the Database appears with the PROVISIONING status and a window shows its password once - save it. The Database is ready in a few minutes, when the status becomes ACTIVE.
| Service | Default port | Features |
|---|---|---|
| PostgreSQL | 5432 | databases with charset and collation, extensions, custom options, Replicas |
| TimescaleDB | 5432 | the same as PostgreSQL |
| MySQL, MariaDB | 3306 | databases with charset and collation, custom options, Replicas |
| MongoDB | 27017 | databases without charset and collation, no Replicas; High Availability runs 3 members |
| Redis | 6379 | no named databases (no DATABASES TAB), users get one access level for the whole instance, backups cover the whole data set |
High Availability (HA) keeps the Database available when its server fails. The platform runs the Database as a primary with one standby copy that receives all changes. If the primary stops responding, the standby is promoted automatically and the connection address stays the same; clients only need to reconnect. MongoDB runs a replica set of three members that elects its primary by itself.
Switch HA on or off with Enable High Availability in the create or edit form. Switching it off removes the standby.
The General block of the details page shows High availability: Yes/No and, when it is on:
| Field | Description |
|---|---|
| Health | Healthy - all members work normally; Suspect, Degraded - a member has a problem, the Database may run without a standby; FailingOver, SwitchingOver - the standby is taking over, after a failure or as a planned operation; SplitBrainRisk - the members cannot agree on the primary, contact customer support |
| Failover delay | how long the platform waits for the primary before promoting the standby |
| Last failover | date and time of the last failover |

Notices. Problems and progress are shown in one box at the top. One notice shows its title and text, as in the picture below; several are merged into one box N notices, most severe first (errors, warnings, information). Notices: Upgrading X → Y (red if the upgrade failed or is blocked); Migration pending approval, in progress, failed or blocked (an older Database is being moved to the current platform image; when it is pending approval, contact customer support to approve or schedule the one restart); Restore failed and Restore failed and was rolled back; Last backup failed; Backup in progress; Load balancer not applied; Extensions not installed, Settings not applied, Charset and collation not applied (server default used), Permissions not applied and Extensions installing after the database restarts; and a notice that databases and users are read-only (see User Management). The reason is written in every notice.

Quick actions (icons next to the title): Start, Stop, Database management (opens the web-based management tool DbGate in a new tab), Reset password, Edit and Delete.
Details area:
| Block | Shows |
|---|---|
| General | ID, creation date, Status, Mode, number of Replicas, High availability and its health, Connection pooler address and mode if the pooler is on |
| Resources | Version (and Running, the version actually running, while it differs, for example during an upgrade), vCPU, Memory, Disk with the used percentage, Charged multiplier with HA |
| Backup | Backing up and Restoring states, Restore phase, Last backup result (hover it for the message), Last run, scheduler Timezone and Next run, S3 link state |
| Connection information | Host (click to copy), S3 Endpoint and S3 Key when backup storage is connected; the secret is never shown, the key is displayed as Configured |
The Usage chart shows CPU and memory load for the last hour.
Tabs. The selected tab is kept in the page address, so a link to it can be shared.
| Tab | Purpose | Shown for |
|---|---|---|
| DATABASES | databases and extensions inside the instance, see Database Management | all except Redis |
| USERS | users and permissions, see User Management | all |
| REPLICAS | copies in other Projects, see Database’s Replicas | all except MongoDB |
| BACKUPS | storage, schedule, backups and restore, see Database’s Backups | all |
| CUSTOM OPTIONS | engine options, see Database’s Custom Options | PostgreSQL, TimescaleDB, MySQL, MariaDB, Redis, MongoDB |
| NETWORK | public and internal endpoints, see Database’s Network | all |
There is no separate scheduler tab: the backup schedule is on the BACKUPS TAB, and an old link with ?tab=scheduler opens it.
Open Edit from the Actions menu on the Databases page or with the Edit icon on the details page, change the fields and click SAVE. The Database is updated in a few minutes.

| Field | Notes |
|---|---|
| Version | the list offers the versions the Database can move to, see Upgrade Database version |
| Port | the port of the Database |
| vCPUs, Memory | resources of the Database (labelled CPU and RAM in the form on the details page) |
| Disk | details page only. The size can only be increased: the field cannot be set below the current size, because a volume cannot shrink. The Database stays online while the disk grows, in most cases without a restart |
| Enable High Availability | see High Availability |
| Connection pooler | details page only, for PostgreSQL and TimescaleDB Databases that already use the pooler (PgBouncer on port 6432, beside each endpoint; the addresses are on the NETWORK TAB). It can be switched off, or its Pooling mode changed: Transaction - clients keep their connection through a switchover or failover, session state (SET, advisory locks, LISTEN, temporary tables) does not outlive a transaction; Session - full session semantics, connections are dropped on a switchover. Switching it off or changing the mode disconnects all clients of the pooler, the form warns you before you save |
The IP ACL is not in the edit form any more: edit it on the NETWORK TAB with Edit allowed CIDRs (see Database’s Network). The databases, extensions, settings and custom options of a Replica follow its primary Database and cannot be edited on the Replica.
Change the version with the Version field of the edit form.
UPGRADING, the notice Upgrading X → Y shows the phase, and Version shows the requested version while Running shows the version that is running now.Create a backup before an upgrade (see Database’s Backups). A major version upgrade cannot be reverted by selecting an older version.
Click the Reset password icon on the details page and confirm. The new password of the main Database user (the DB user from the create form) is shown once - save it. The icon is active for running Databases (ACTIVE or DEGRADED) in Standalone or Primary mode. It is not available for Replicas and for Databases whose users are managed inside the engine. Passwords of other users are reset on the USERS TAB.
STOPPED Database.Deleting a Database permanently removes its data. This action cannot be undone.