Highlights
- Run 8 self-hosted models on GreenNode MaaS, from GLM-5.3 to embedding and rerank models, processed on GreenNode infrastructure in Vietnam.
- Deploy Redis Cluster in Hanoi across two Availability Zones, HAN 1a and HAN 1b, for high availability closer to workloads in northern Vietnam.
- Resize Redis Cluster instances up or down from the Portal, with automatic rollback if a resize fails.
- Schedule recurring jobs inside PostgreSQL Cluster with pg_cron, no external scheduler needed.
- Find any subnet in a project by network, zone, name, CIDR or ID, and search Servers and Networks by ID.
October's updates focus on flexibility: more places to run your cache, more control over how it scales, and less work to keep databases and networks in order.
This month, GreenNode brings vDB Redis Cluster to the Hanoi region, adds instance resizing for Redis Cluster, and supports the pg_cron extension on PostgreSQL Cluster. On vServer, you can now view and search every subnet in a project from one place.
vDB: More regions and more control for Redis Cluster
Redis Cluster now available in Hanoi
Redis Cluster (MemoryStore) is now available in the Hanoi (HAN) region, across two Availability Zones:
- HAN 1a and HAN 1b, so cluster nodes can be spread across zones for high availability.
Teams running applications in northern Vietnam can now place their cache in the same region as their workloads. That means lower latency for read-heavy services and an HA setup without routing traffic to another region.
Resize Redis Cluster instances (scale up/down)
You can now change the instance type of a Redis Cluster directly in the Portal:
- Pick a new package to scale up or down.
- Quota check: the system confirms you have enough quota before it starts.
- Automatic rollback: if the resize fails, the cluster returns to its previous configuration.
This feature is currently available in the HAN region.
When traffic grows or drops, you can match cache capacity to demand without rebuilding the cluster, and without risking a half-finished change.
vDB: pg_cron support for PostgreSQL Cluster
PostgreSQL Cluster now supports the pg_cron extension, which lets you schedule recurring jobs inside the database itself. Configure it through the cron.database_name parameter.
Routine tasks such as cleaning old data, refreshing materialized views or running periodic reports can now run on a schedule without setting up and maintaining a separate scheduler.
vServer: View and search subnets by project
Subnets used to be viewable only within each network. Now you can list every subnet in a project at once and filter by network, zone, name, CIDR range or subnet ID, with partial matching for name and CIDR. The Server and Network pages also add search by ID.
This applies to customer-created subnets. It is live in the HCM region, and the HAN region is planned for the second week of November 2026.
For teams managing many networks, this cuts the time spent clicking through each one to find a subnet or check which CIDR ranges are already in use.
MaaS: 8 self-hosted models now running in Vietnam
Since October 31, 2026, GreenNode MaaS serves 8 self-hosted models on GreenNode infrastructure in Vietnam. Requests are processed on GreenNode GPUs in Vietnam, called through an OpenAI-compatible API and billed per token, with no GPUs to manage.
| Type | Models | Best for |
|---|---|---|
| Large text model | GLM-5.3 (Z.ai) | Complex reasoning, coding, multi-step agents |
| Multimodal everyday model | GLM-5.3 Flash (Z.ai) | High-volume tasks: customer chatbots, internal assistants, document extraction |
| Small models | Gemma 4 31B (Google DeepMind), Qwen3.8-27B (Qwen) | Classification, data extraction, batch jobs at lower cost |
| Embedding | bge-m3 (BAAI), Qwen3-Embedding-8B (Qwen) | Semantic search, the retrieval step of RAG |
| Rerank | bge-reranker-v2-m3 (BAAI), Qwen3-Reranker-8B (Qwen) | Ranking retrieved documents before they reach the prompt |
For businesses that must comply with rules on processing and transferring personal data, such as Vietnam's Personal Data Protection Law in effect since January 1, 2026, self-hosted models make it clear where requests are processed. Third-party models remain available on the same platform when a commercial model is needed.
See the available models list for model IDs and serving details, then test each model on the MaaS Playground with your own prompts.
What these updates mean
Together, October's releases give teams more room to grow without adding operational work.
- Closer to your users: Redis Cluster in Hanoi keeps cache and application in the same region.
- Scale on demand: resize Redis Cluster as traffic changes, with rollback if anything goes wrong.
- Fewer moving parts: run scheduled database jobs with pg_cron instead of an external scheduler.
- Faster network management: find any subnet in a project in one search.
The result is less time on infrastructure housekeeping and more time on the applications that run on it.
Technical documentation
Explore the following documentation to get started with the October 2026 updates:
- vDB feature overview - check Redis Cluster availability by region.
- Change Redis Cluster instance type - resize a cluster step by step.
- Supported PostgreSQL Cluster extensions - see the full extension list, including pg_cron.
- Configure pg_cron parameters - set cron.database_name for your cluster.
- 2026 product updates - follow all release notes, including subnet search rollout to HAN.
Get started
Log in to the GreenNode Portal to try the new features today.
Need help planning a deployment? Contact the GreenNode team for consultation and implementation support.
Frequently Asked Questions
Which Availability Zones does Redis Cluster support in Hanoi?
Redis Cluster in the Hanoi region runs across two Availability Zones: HAN 1a and HAN 1b.
Can I resize a Redis Cluster in every region?
Not yet. Instance resizing for Redis Cluster is currently available in the HAN region.
What happens if a Redis Cluster resize fails?
The system automatically rolls the cluster back to its previous configuration, so the cluster is not left in a partial state.
How do I enable pg_cron on PostgreSQL Cluster?
pg_cron is a supported extension on PostgreSQL Cluster. Configure it through the cron.database_name parameter, as described in the pg_cron configuration guide.
Which subnets appear in the project-wide subnet view?
The view lists customer-created subnets in the project. You can filter by network, zone, name, CIDR range or subnet ID.
When will project-wide subnet search reach the HAN region?
It is live in HCM now. The HAN release is planned for the second week of November 2026.
