Scale a Hyperscale (Citus) server group

APPLIES TO: Azure Database for PostgreSQL - Hyperscale (Citus)

Azure Database for PostgreSQL - Hyperscale (Citus) provides self-service scaling to deal with increased load. The Azure portal makes it easy to add new worker nodes, and to increase the vCores of existing nodes. Adding nodes causes no downtime, and even moving shards to the new nodes (called shard rebalancing) happens without interrupting queries.

Add worker nodes

To add nodes, go to the Compute + storage tab in your Hyperscale (Citus) server group. Dragging the slider for Worker node count changes the value.

Note

A Hyperscale (Citus) server group created with the basic tier has no workers. Increasing the worker count automatically graduates the server group to the standard tier. After graduating a server group to the standard tier, you can't downgrade it back to the basic tier.

Click the Save button to make the changed value take effect.

Note

Once increased and saved, the number of worker nodes cannot be decreased using the slider.

Note

To take advantage of newly added nodes you must rebalance distributed table shards, which means moving some shards from existing nodes to the new ones. Rebalancing can work in the background, and requires no downtime.

Increase or decrease vCores on nodes

In addition to adding new nodes, you can increase the capabilities of existing nodes. Adjusting compute capacity up and down can be useful for performance experiments, and short- or long-term changes to traffic demands.

To change the vCores for all worker nodes, adjust the vCores slider under Configuration (per worker node). The coordinator node's vCores can be adjusted independently. Adjust the vCores slider under Configuration (coordinator node).

Note

There is a vCore quota per Azure subscription per region. The default quota should be more than enough to experiment with Hyperscale (Citus). If you need more vCores for a region in your subscription, see how to adjust compute quotas.

Increase storage on nodes

In addition to adding new nodes, you can increase the disk space of existing nodes. Increasing disk space can allow you to do more with existing worker nodes before needing to add more worker nodes.

To change the storage for all worker nodes, adjust the storage slider under Configuration (per worker node). The coordinator node's storage can be adjusted independently. Adjust the storage slider under Configuration (coordinator node).

Note

Once increased and saved, the storage per node cannot be decreased using the slider.

Next steps

• Learn more about server group performance options.
• Rebalance distributed table shards so that all worker nodes can participate in parallel queries
• See the sizes of distributed tables, and other useful diagnostic queries.