rds.replication.setMake a cluster synchronous or asynchronous (live, no restart), and/or turn logical replication on or off (with a restart)
Synchronous is not strict: with no standby streaming, the leader keeps accepting writes and the health check reports sync-degraded until one is back. logicalReplication sets wal_level = logical; a cluster restarts member by member with one switchover, a single node restarts once (writes pause for the restart). Turning it off is refused while any logical slot exists.
Request syntax
PATCH /rds/api/instances/{id}/replication
{
"replicationMode": "enum",
"logicalReplication": "boolean"
}Path parameters
idRequiredstring
Constraints: 3-64 chars, /^(?!ewdoc-)[a-z0-9-]+$/
Request body
replicationModeOptionalenum
Cluster only, applied live. sync: a commit waits for one standby, so a failover loses nothing acknowledged; async: it does not wait. Omitted: unchanged
One of:
async,synclogicalReplicationOptionalboolean
wal_level = logical for CDC and publications, applied with a restart; omitted: unchanged
Response
200 with instance (the database, as `get` describes it without `placement`), applied, replication ({ mode, state, syncStandbys }, null when replicationMode was not sent) and logicalReplication ({ enabled, applying }).
A JSON response is wrapped as { "success": true, "data": … }; the CLI prints the whole body with --json.
Fields of data
instanceAlwaysobject
The databaseinstance.idAlwaysstring
Database id (ewdb-…)instance.orgIdAlwaysstring
Organization that owns the databaseinstance.identifierAlwaysstring
Name of the database, unique within the organizationinstance.engineAlwaysone of postgres, docdb
Database engine; postgres for every managed PostgreSQL databaseinstance.engineVersionAlwaysstring
PostgreSQL version the database is on, a pinned minor such as 17.11instance.instanceClassAlwaysstring
Instance type of every node, e.g. performance-2instance.regionIdAlwaysstring
Region the database is in, e.g. de1instance.availabilityZoneIdAlwaysstring, may be null
Zone asked for at create (e.g. de1b); null: the region's default zone, or spread over the zones for a clusterinstance.vpcIdAlwaysstring, may be null
VPC the database joins; null: noneinstance.subnetIdAlwaysstring, may be null
Subnet of that VPC its addresses come from; null: the VPC's whole rangeinstance.dbNameAlwaysstring, may be null
Database created inside the server at launchinstance.masterUsernameAlwaysstring
Name of the master role, whose password is filed in Secretsinstance.allocatedStorageGibAlwaysnumber
Size of each node's data volume, GiBinstance.maxAllocatedStorageGibAlwaysnumber, may be null
Autoscaling ceiling, GiB; null: automatic growth is offinstance.storageTypeAlwaysstring
Storage class label; informationalinstance.iopsAlwaysnumber, may be null
Provisioned IOPS; not offered, so nullinstance.multiAzAlwaysboolean
Not used: zonePlacement says how a cluster is spread over zonesinstance.publiclyAccessibleAlwaysboolean
Whether the database answers on the internet; false: only from its VPCinstance.topologyAlwaysone of single, cluster
single: one node; cluster: three nodes with automatic failoverinstance.nodeCountAlwaysnumber
Core nodes: 1 for single, 3 for a cluster (readers not counted)instance.masterSecretIdAlwaysstring, may be null
Secret holding the master password; null until it is filedinstance.loadBalancerIdAlwaysstring, may be null
Load balancer in front of a cluster, on the platform's own fleet; null otherwiseinstance.endpointKindAlwaysone of hetzner-lb, fleet, may be null
What a cluster's endpoint is: fleet (the platform's own load balancer) or hetzner-lb (the provider's); null on a single nodeinstance.hetznerLoadBalancerIdAlwaysnumber, may be null
The provider's id of a cluster's load balancer; null when there is noneinstance.hetznerNetworkIdAlwaysnumber, may be null
The provider's id of the private network a cluster's members talk over; null on a single nodeinstance.hetznerFirewallIdAlwaysnumber, may be null
The provider's id of the database's firewall; null before it existsinstance.reconcileErrorAlwaysstring, may be null
Why the last provisioning or teardown step failed; null when nothing failedinstance.capacityFallbackAlwaysobject
What the create could fall back to when its type was sold outinstance.capacityFallback.otherZonesAlwaysboolean
A sold-out zone may be replaced by another zone of the regioninstance.capacityFallback.equivalentTypesAlwaysboolean
A sold-out instance type may be replaced by an equivalent oneinstance.onStockOutAlwaysone of fail, wait
Sold out with no fallback: fail the create, or wait for stockinstance.capacityWaitUntilAlwaysstring, may be null
While waiting for stock: when the create gives up; null otherwiseinstance.requestedInstanceClassAlwaysstring, may be null
The instance type asked for, when a stock fallback bought another; null otherwiseinstance.statusAlwaysone of available, creating, modifying, backing-up, deleting, failed, stopped, starting, stopping, rebooting, maintenance, waiting-for-capacity, retained
State of the databaseinstance.retainedUntilAlwaysstring, may be null
For a retained database: when its backups are removed; null otherwiseinstance.endpointAddressAlwaysstring, may be null
Address clients connect to; null until the database is reachableinstance.endpointPortAlwaysnumber, may be null
Direct PostgreSQL port of the endpoint (5432); the pooler is on 6432instance.privateEndpointAddressAlwaysstring, may be null
The database's address inside its VPC, named in its certificate; null: use endpointAddressinstance.parameterGroupIdAlwaysstring, may be null
Not used; parameters holds your valuesinstance.backupRetentionDaysAlwaysnumber
Days backups are kept; 0: scheduled backups are offinstance.backupsEnabledAlwaysboolean
Whether scheduled backups are takeninstance.storageEncryptedAlwaysboolean
Whether every node's data volume is encrypted at restinstance.storageKmsKeyIdAlwaysstring, may be null
KMS key that wraps every node volume's key; null: the platform keyinstance.backupAgentSeenAtAlwaysstring, may be null
When a node last asked for backup work; null: neverinstance.backupPollerVersionAlwaysnumber, may be null
Version of the backup service on the nodes; 3 and up keep a backup taken by hand until it is deleted; null: not reportedinstance.lastBackupAtAlwaysstring, may be null
When the newest backup finished; null: neverinstance.backupWindowAlwaysstring, may be null
When the daily backup runs, UTC HH:MM-HH:MM (only the start is used)instance.maintenanceWindowAlwaysstring, may be null
Maintenance window, UTC ddd:hh24:mi-ddd:hh24:miinstance.deletionProtectionAlwaysboolean
While true, delete is refusedinstance.performanceInsightsEnabledAlwaysboolean
Not used; query insights are always oninstance.autoMinorVersionUpgradeAlwaysboolean
Whether new minor versions are applied in the maintenance windowinstance.parametersAlwaysmap
Your PostgreSQL parameter values by name; parameters not named keep their defaultinstance.caCertIdentifierAlwaysstring, may be null
Name of the CA that signs the server certificateinstance.tlsSerialAlwaysstring, may be null
Serial of the server certificate; null: the database has no certificateinstance.tlsNotAfterAlwaysstring, may be null
When the server certificate expiresinstance.tlsNamesInCertAlwaysboolean
Whether the certificate names the database's DNS endpoint namesinstance.tlsRequiredAlwaysboolean
Whether clients that do not use TLS are refusedinstance.replicationModeAlwaysone of async, sync, may be null
sync: a commit waits for one standby; async: it does not; null on a single nodeinstance.zonePlacementAlwaysone of spread, single, may be null
spread: members over the zones; single: all in one zone; null on a single nodeinstance.logicalReplicationAlwaysboolean
Whether wal_level is logical, for CDC and publicationsinstance.branchPoolOfAlwaysstring, may be null
Set on the machine that serves the branches of another database: that database; null otherwiseinstance.branchSettingsAlwaysunknown
Branch settings as stored (branch-settings reads them); null: the defaultsinstance.launchedAtAlwaysstring, may be null
When the database first became availableinstance.tagsAlwaysstring, may be null
Kept for compatibility; the tags API holds the tags of the databaseinstance.createdAtAlwaysstring
When the database was createdinstance.updatedAtAlwaysstring
When the database record last changedinstance.encryptionAlwaysobject
What is encrypted at restinstance.encryption.storageAlwaysone of luks2, none
luks2: every node's data volume is encrypted with a per-volume key from KMS; none: created before disk encryptioninstance.encryption.backupsAlwaysone of aes-256-cbc, none
aes-256-cbc: backups are encrypted on the database machine before upload; none: created before backup encryptionappliedAlwaysboolean
The cluster took the replication mode (always true when it was not sent)replicationAlwaysobject, may be null
Replication now; null when replicationMode was not sentreplication.modeAlwaysone of async, sync
The mode asked forreplication.stateAlwaysone of ok, degraded, async, unknown
ok: sync and a standby acknowledges commits; degraded: sync but none does; async; unknown: the cluster did not answerreplication.syncStandbysAlwaysarray of string
Members a commit waits for (any one of them)logicalReplicationAlwaysobject
Logical replicationlogicalReplication.enabledAlwaysboolean
wal_level logical is asked forlogicalReplication.applyingAlwaysboolean
A restart that applies it was started by this call
Checked against the handlers: every route test validates their answers against this schema.
Errors
- 400The request does not pass the schema above. The body is { "error": …, "issues": […] }, one issue per field.
- 401No credentials, or an access key that is unknown, revoked or expired.
- 403The caller's policies do not allow rds:ModifyDBInstance on this resource. The body names the action and the resource, and the refusal is recorded in the audit log.
- 404The resource does not exist in your organization.
Examples
The required parameters only, with example values. Replace id and the other placeholders with your own.
dcloud rds replication set idcurl -X PATCH "https://console.platform-development.deutschecloud.app/rds/api/instances/id/replication" \
-H "Authorization: EWS-ACCESS-KEY $ACCESS_KEY" \
-H "Content-Type: application/json" \
-d '{}'Generated from the API description, the same one the API checks requests against and the CLI is built from. MCP tool: dcloud_rds_replication_set.