Instances
The Instances tab displays all configured instances for your PostgreSQL database service in a table, with per-instance details and actions including switchover, DR drills, and replica creation.
What instance topologies does Tessell support for PostgreSQL?
Single instance
One instance in a given region and availability zone.
Multi-AZ (HA)
Two instances — one primary and one synchronous replica — in the same region (same or different AZ). Both instances are always homogeneous (infrastructure, parameters, storage, IOPS). All planned operations run in rolling mode to minimise downtime.
Disaster Recovery (DR)
One or more async replica instances added in the same or a different region. No automatic failover. Can be added to both single-instance and Multi-AZ services.
Read Replica (RR)
An async replica opened in read-only mode. Use for read-only query offloading.
How do I add a High Availability (HA) replica?
You can convert a single-instance (SI) service to HA, even if it already has DR or Read Replica instances attached — the existing DR/RR instances remain intact and continue replicating from the new HA primary after conversion.
Prerequisites: Open bi-directional network communication between the primary and HA subnets at the subnet level for the primary database port (e.g., 5432 for PostgreSQL) and port 1599. This must cover the entire CIDR blocks of the subnets. NSG rules must be applied at the subnet level, not the VM level — VM-level NSG is managed by Tessell.
Steps to add an HA instance
Click Create Replica (top-right of the Instances tab) → select High Availability.
In the Add High Availability Instance window:
Select an Availability Zone from the dropdown. Region and VNet stay the same as the primary and cannot be changed.
Select an available Private Subnet from the dropdown.
Review the HA properties inherited from the primary — Compute shape, vCPUs, Storage, Encryption Key, IOPS, Throughput. These cannot be changed for HA.
Click Create.
Note:
Brief downtime (typically 30–60 seconds) occurs when the primary PostgreSQL service is stopped and restarted under Patroni management.
Tessell automatically:
Runs pre-conversion validation on the SI service.
Provisions the secondary compute (creates the secondary VM).
Installs the required software (PostgreSQL, Patroni, and etcd).
Sets up the etcd cluster (forms quorum with the primary, standby, and controller).
Configures the primary — converts standalone PostgreSQL into a Patroni-managed cluster.
Synchronises the standby — clones data from the primary using
pg_basebackup.
How do I add a DR or Read Replica instance?
You must open bi-directional network communication between the primary and the DR/RR subnets:
Primary database port (e.g., 5432 for PostgreSQL, or your specific database port)
Port 1599
Applied at the subnet level (entire CIDR blocks), not VM level — NSG rules at the VM level are managed by Tessell.
Steps to add a DR or Read Replica instance
Click Create Replica (top-right of the Instances tab) → select Disaster Recovery or Read Only Replica.
Provide the following details:
Instance Name — alphanumeric characters only.
Region — target region.
Availability Zone — target AZ.
VPC — target VPC.
Private Subnet — target private subnet.
Click Add Instance.
Tessell automatically:
Submits the request to add the replica.
Creates the compute instance(s).
Prepares the primary database to accept the new replica.
Configures the replica with primary database information.
Sets up the Availability Machine for the new instance.
As an alternative, use the Code button to view, copy, or download the API call for adding an instance in Shell, Python, Go, Java, JavaScript, or PowerShell.
What columns appear in the Instances table?
Name
Auto-generated name: default-node-0 (primary), default-node-1 (typically the HA replica configured for automatic failover), <instance name>-node-<digit> (RR/DR, digit incrementing per node added).
Status
Current instance state (see table below).
Role
Primary, HA Replica (Automatic Failover), Read Replica (RR), or Disaster Recovery (DR). Shows the server name below the role.
Server
Server where the instance is deployed. Click the server name to open its home page in the Servers app.
Region/Zone
Cloud provider and region hosting the instance.
Parameter Profile
Parameter profile applied at provisioning.
Compute
Compute shape with vCPU and memory details below.
Created
Date and time the instance was created.
Storage
Storage size; click See details for storage provider, throughput, and IOPS.
Connection Endpoint
Hostname; click See details for Service URL, Host, Port, and Username.
Instance status reference
UP
Instance is running and accepting connections.
Resize compute/storage, update parameter profile, delete (RR/DR only)
DOWN
Instance is down.
Raise a Tessell support ticket to have it revived
STOPPED
Instance is completely stopped.
Start the service
CREATING
Provisioning is in progress, or a new instance is being added to the service.
No actions
DELETING
Deletion in progress (RR/DR only).
No actions
REBUILDING
Compute or storage resize in progress.
No actions
REBUILD_FAILED
Resize was triggered and did not complete.
All actions
PATCHING
A patch is being applied.
No actions
PATCH_FAILED
Patching failed.
All actions
UPDATING_PARAMETER_PROFILE
Parameter profile update in progress.
No actions
PARAMETER_PROFILE_UPDATE_FAILED
Parameter profile update failed.
All actions
Note: Using the ellipsis icon on individual instance rows, you can create, update, or delete private links for DR and RR nodes separately. Creation of private links for DR and RR nodes is currently supported for AWS cloud only.
How do I manually trigger a switchover to a replica?
Switchover promotes a replica (HA or DR) to become the new primary.
Note: Switching over to an instance may cause some downtime for the service. If an HA replica is chosen as the target, the switchover is graceful (planned and synchronised with no data loss) and a switchback is allowed.
Steps to initiate a switchover
Click Switch over on the Instances tab.
In the Switch over to an instance window, select the target replica instance from the list of available replicas (shown with their names and statuses).
Optionally add comments.
Select the I understand the consequences checkbox.
Click Switch.
Monitor progress by clicking Submitting Request below the service name. Once the switchover completes, the service returns to Ready status and Tessell metadata is updated.
What is the difference between a DR Drill and an actual disaster recovery switchover?
When switching over to a DR instance, Tessell offers two paths, controlled by the DR Drill flag:
DR Drill (Graceful Switchover)
Yes
Attempts a graceful switchover: the DR instance is checked for replication sync with the primary, the current primary is demoted in a controlled manner, and the DR instance is promoted with data consistency. If the DR instance is out of sync, the switchover fails to prevent data inconsistency. Used for DR testing and compliance drills — switch back to the original primary is possible after testing.
Actual Disaster Recovery (Forceful Switchover)
No
Treats the scenario as a real disaster. Tessell forcefully promotes the DR instance to primary and brings down the original primary. The DR node requires a manual rebuild to restore replication consistency. Use only for an actual failover.
Note: Ensure you select the DR Drill option when performing a planned DR test. Initiating a switchover without the DR Drill option enabled treats the operation as a real disaster recovery and brings down the original primary.
Figure 1 — Instances tab
Last updated
Was this helpful?