2026

August 2026

vDB – PostgreSQL Cluster in HAN region

PostgreSQL Cluster (RDS) is now available in the HAN-01 region, bringing High Availability PostgreSQL to Hanoi with availability zones HAN01-1A and HAN01-1B.

Deploy 1 Writer + N Readers clusters (2–10 nodes) with automatic failover in HAN-01.
Supported AZs: HAN01-1A, HAN01-1B — alongside the existing HCM region.
Learn more at PostgreSQL Cluster.

GreenNode MaaS – Model Catalog & Pricing Update

GreenNode is updating the MaaS model catalog into two groups — models self-hosted directly by GreenNode and third-party models under official contracts — along with more competitive pricing on many models. The Portal now also shows a model type label (Self-host / Partner) right on the model list, making it easy to tell models apart and pick the right one.

Current models are discontinued and the new catalog is published on August 3, 2026, with an automatic 30-day extension until September 2, 2026 before old-model requests start returning errors.
Compare your current models against the new catalog; only models no longer listed require choosing a replacement.
Learn more at AI Stack — Release Notes.

GreenNode AI Gateway — Architecture Upgrade, Optimized Performance

GreenNode has upgraded the AI Gateway architecture to better handle growing traffic from MaaS and Token Plan.

Low Latency: Reduced latency in API key authentication and rate limiting.
Scalability: No longer limited by memory — serves more models and more users.
High Performance: Handles Token Plan traffic reliably.
Learn more at AI Stack — Release Notes.

June 2026

vMonitor – Added vStorage metrics for regions HCM-04, HAN-02

vMonitor adds a metric set for vStorage in the HCM-04 and HAN-02 regions, supporting monitoring of traffic, requests, latency and capacity per bucket. See the updated metric list at List of vStorage metrics.

Cost Explorer

Cost Explorer has received a comprehensive upgrade, allowing you to analyze service usage costs across multiple data dimensions, at different time aggregation levels, and drill down to individual resources — helping customers understand where costs come from and optimize their resource usage.

Flexible aggregation levels: View costs at four time resolutions — Hourly, Daily, Weekly, Monthly — within the current month.
Drill-down to individual resources: The Resource ID view lists the cost of each resource (total, daily average, start date); click a row to see the resource details.
Three cost views: Aggregate costs by Product, by Resource Type (with a "Top resource type" card and a "% of total" column), or by Resource ID.

VKS - Kubernetes 1.31 and 1.32 Support

VKS (VNG Kubernetes Engine) officially supports Kubernetes versions 1.31 and 1.32, allowing users to create new clusters or upgrade existing clusters to the latest versions.

Brings improvements in stability, security, performance, and compatibility with the Kubernetes ecosystem.
Helps users keep their environments aligned with community-supported Kubernetes versions.
Learn more at Upgrade Kubernetes Version.

May 2026

VKS - Cluster-level Auto Healing

VKS (VNG Kubernetes Engine) now supports configuring Auto Healing directly at the cluster level, giving you control over unhealthy node detection thresholds and self-recovery behavior tailored to your workload type. When a node remains unhealthy long enough, the system automatically deletes the faulty node and provisions a replacement — no manual intervention required.

Flexible enable/disable; configure Timeout, Max Unhealthy, and Unhealthy Range via the Portal.
Built-in cluster protection: automatically halts replacements when faulty nodes exceed the threshold, preventing mass deletion during widespread infrastructure incidents.
Configuration changes take effect immediately — no cluster or node restart required.
Learn more at Auto Healing.
API documentation for creating and managing Auto Healing via API: VKS API.
Terraform documentation for deploying Auto Healing with Infrastructure as Code: VKS Terraform.

April 2026

VLB - Access Control List (ACL) on Listener

The ACL (Access Control List) feature for Load Balancer has been launched as a comprehensive upgrade over the previous IP Whitelist. Instead of only allowing configuration of an Allowed CIDRs list, ACL adds the ability to actively block traffic via Dropped CIDRs and define Default Actions when no rules match — enabling stricter, more environment-appropriate access policies.

Learn more at: ACL NLB, ACL ALB.

VLB - Auto Scaling – Support for editing configuration after initialization

VLB now supports editing Auto Scaling configuration after creation, through the Configure AutoScale function on the detail page. Specifically, the following three fields are now open for editing:

Min size – The minimum number of instances in the cluster (Min: 1, Max: 15).
Max size – The maximum number of instances in the cluster (Min: 1, Max: 15).
Availability zones – The availability zones and corresponding subnets used for the Load Balancer cluster.

Previously, these parameters could only be set at the time of initialization and could not be changed afterward. This update allows flexible adjustment of scaling limits and expansion of availability zones without needing to recreate the Load Balancer.

Learn more at: Auto Scale.

VKS - Kubeconfig with Flexible Certificate Validity

VKS (VNG Kubernetes Engine) now allows you to actively choose the certificate validity period when downloading kubeconfig, providing better security control. Supported validity options: 30 days, 90 days, 365 days — suitable for different environments (testing, production, long-term).

View the certificate expiry date directly on the Portal before downloading.
Automatic certificate renewal when eligible; manual renewal supported when needed.
Learn more at Kubeconfig.

March 2026

New display for wallet credits balance and service pricing will be updated to show prices before VAT

Changes: VAT will be excluded from the pricing display on the Portal. Specifically:

Wallet Credit balance will be displayed based on the pre-VAT value;
Service unit prices will also be displayed based on the pre-VAT rate.

Principles of Credit Display After the Changes:

This update only changes the display method and does not affect the actual usable value or your benefits. Details are as follows:

Item

Before Update (Incl. VAT)

After Update (Excl. VAT)

Notes

Wallet balance

110 Credits

100 Credits

Actual value unchanged

Price per service

110 Credits

100 Credits

Actual value unchanged

Purchasable quantity

01 Service

No change

Credit Top-up and Usage Rules

Criteria

Before 01/03/2026

From 01/03/2026 onwards

Top-up rule & Credits received

The top-up amount includes 10% VAT; Credits received = top-up amount

The top-up amount still includes 10% VAT Credits received = top-up amount ÷ 110%

The rule of Credit deduction when creating resources

Credits are deducted based on the VAT-included price (10% VAT)

Credits are deducted based on the pre-VAT price

February 2026

vServer - Allows selecting MTU when creating a VPC.

Virtual Private Cloud (VPC) networks use a default MTU value of 1500 bytes starting from February (the old value was 1450).

You can set the MTU of a VPC network to any value in the range from 1,450 bytes to 8,950 bytes.

vServer - Allows exporting the server list.

Exporting the server list supports formats such as xls, json, and csv.

vServer - Adds Windows Server 2025 image with license.

Allows customers to provision a server with Windows Server 2025 OS, including a license.

vStorage – Update on CopyObject API Limitation with Encryption.

The vStorage (Object Storage) CopyObject API is not supported for buckets or objects with Server-Side Encryption enabled (SSE-S3 or SSE-C) and will return 501 – Not Implemented.

VKS - Terraform Provider for VNG Cloud Upgrades

VKS (VNG Kubernetes Engine) has just released its latest update, bringing several new upgrades to users. Below are the key highlights of this release:

New Upgrades:

Terraform Provider for VNG Cloud v1.3.7: Updated the cluster creation workflow to use the POC provisioning API. For more information, see here.
Terraform Provider for VNG Cloud v1.3.8: Fixed an issue where node group labels and taints were missing after terraform import. Node groups imported from the Portal now have labels and taints correctly reflected in Terraform state. For more information, see here.

January 2026

Update on Expired Resource Renewal Mechanism on the Portal

To optimize resource management and provide additional time for expired resources to be retained on GreenNode—allowing Customers more time to complete renewal procedures—we are updating the renewal mechanism for expired resources on the Portal (extended from 7 days to 30 days). This change applies from the moment a resource reaches its expiration date within the usage cycle. Key updates to the renewal policy are as follows:

After expiration, resources will transition through the following stages:

Grace Period: From Day 1 to Day 15 after the expiration date. During this period, the service remains active and data is fully retained.
Suspension: From Day 15 to Day 30 after the expiration date. During this period, the service is suspended, but data is still retained in the system.
Release and Deletion: After 30 days from the expiration date. At this stage, the service is permanently terminated, and all associated data is irreversibly deleted from the system.

Throughout the Grace Period and Suspension stages (i.e., from Day 1 up to before Day 30 after expiration), resources can be manually renewed to continue using the service.

This renewal policy will take effect from January 8, 2026, and applies to prepaid wallet-based resources with expiration dates occurring after this time.

For more details about the updated renewal mechanism and policy, please refer here.

Aug 2026

vMonitor — vMonitor Datasource Plugin for Grafana

The vMonitor Datasource Plugin connects Grafana to the vMonitor Platform so you can query metrics and logs, build dashboards, and drive Grafana alerting directly from vMonitor data.

Queries metrics and logs from vMonitor, with metric statistics, dimension filters, group by, and functions including Rank, Rollup, Rate, Timeshift, and Reduce.
Supports Expression mode for arithmetic across queries (e.g. A / B * 100) and Grafana alerting and notifications on vMonitor metrics and logs.
Ships bundled dashboards for Host, vServer, vLB, vDB, vStorage, and vBackup, plus template variables for metrics, dimensions, and infrastructure resources.
Requires Grafana 12.3.0+ and a GreenNode IAM Service Account key pair; unsigned builds must be allowlisted — download from GitHub releases.
Learn more at Install and use the vMonitor Datasource plugin for Grafana.

GreenNode AgentBase — CLI & MCP

AgentBase adds two ways to work alongside the Portal — the command line and AI assistants:

GreenNode CLI: the grn agentbase command group manages the full agent lifecycle (Identity, Runtime, Memory, MCP Gateway, Policy, Container Registry) and deploys a whole agent from a single manifest — ideal for moving agent deployment into automation scripts and CI/CD.
AgentBase MCP: lets AI assistants (Claude Desktop, Claude Code, Cursor, Windsurf...) operate AgentBase through natural language via 3 meta-tools, with both local (stdio) and remote (HTTP) connections.
Learn more at AI Stack — Release Notes.

GreenNode AI Stack — Token Plan

Token Plan lets you lock in your monthly AI budget with a 30-day prepaid package that has a fixed token/request limit per model — instead of paying per token actually used like PAYG.

Access via a subscription-key, fully separate from PAYG API Keys.
Browse and buy packages directly under the Token Plan group on the Portal.
Suited for stable daily usage (e.g. coding assistants).
Learn more at AI Stack — Release Notes.

VKS - VKS MCP Server

Connect AI assistants (Claude, Cursor, VS Code…) straight to VKS over the Model Context Protocol — manage Clusters, Node Groups and the Kubernetes resources inside a cluster in plain language.

Remote (hosted): use the hosted endpoint right away and sign in with your GreenNode IAM user in the browser (OAuth 2.1) — nothing to install, no secret in any config file. See Configure Remote MCP.
Local (stdio): run the server on your own machine, sharing the ~/.greennode credentials with the GreenNode CLI. See Configure Local MCP.
See VKS MCP Server for details.

vServer - Scheduled O&M (Beta): schedule automated VM operations

vServer introduces Operation and Maintenance (O&M) — Scheduled Task in Beta, letting you schedule START, STOP and REBOOT operations on your virtual machines at a fixed time instead of performing them manually every day.

Schedule runs daily at the hour and minute you choose, within an effective window (start date — end date).
Select targets by VM list or by tag (tag_key/tag_value) — VMs tagged later are picked up automatically with no need to update the task.
Full task lifecycle management: create, update, duplicate, cancel, delete, plus Run now to execute without waiting for the schedule.
Execution history down to each VM: success/failure status with a specific error message, filterable by status and trigger type.
One-time Activate: the system provisions a dedicated Service Account per user with least-privilege permissions, so scheduled operations run securely under your own identity.
Available in HCM-03 and HAN-01; default quota of 20 tasks per account.
Learn more at Scheduled O&M.

Jul 2026

IAM - IP Whitelist

IAM introduces IP Whitelist, letting you restrict sign-in to your GreenNode account to approved networks only — an extra layer of protection at the sign-in stage, alongside password policy and MFA.

Declare up to 10 IP addresses or CIDR ranges (IPv4 and IPv6) per rule.
Apply rules by Scope: Root user, IAM user and Identity provider; Service Accounts are not affected.
The Enabled toggle turns a rule on or off instantly without deleting its configuration — handy when you need to lift a restriction temporarily.
The Enabled immediately option lets you create a rule now and activate it later, reducing the risk of locking yourself out.
Learn more at Configure IP Whitelist.

VKS - GreenNode CLI

The GreenNode CLI (the grn command) lets you manage the full lifecycle of VKS Clusters and Node Groups directly from your terminal — ideal for fast, repeatable operations and automation scripts.

Learn more at Manage VKS with the GreenNode CLI.

GreenNode AgentBase — Security Runtime & MCP Connectors

GreenNode AgentBase adds new security and integration capabilities for AI Agents:

Security Runtime: Configure IP Access Control (allowed source IP CIDR ranges) and Inbound Identity (IAM Permissions, JWT, or No authorization) directly when creating a Runtime.
MCP Connectors: Connect an agent to external services — GitHub, Slack, Microsoft 365... — in minutes via a prebuilt catalog, without building an MCP server or hand-coding OAuth.
Learn more at AI Stack — Release Notes.

Jun 2026

vDB - Redis Cluster (Non-sharding) for MemoryStore (MDS)

GreenNode introduces Redis Cluster (Non-sharding) for the vDB MemoryStore (MDS) service, bringing a High Availability architecture and comprehensive data protection for Redis workloads in production.

1 Master – up to 9 Replicas: continuous replication, serving read scaling and ready for failover.
Automatic Failover: automatically promotes a Replica to the new Master when the Master fails.
Zero-downtime replica scaling: add/remove one node at a time, ensuring sync completes before each step.
vBackup integration: easily configure Auto/Manual backups and restore from the Backup Center.
Supports Redis version 7.x.
Learn more at Redis Cluster.

VKS - Resource Tag

Resource Tag lets you manage a set of key/value labels across all resources — including Servers and Volumes — that belong to a Node Group.

Learn more at Resource Tag.
VKS API Docs: VKS API.
Terraform documentation for creating a node group as Infrastructure as Code: VKS Terraform.

May 2026

vMonitor - Export vLB access logs to vDB

vMonitor now lets you export vLB access logs to your own vDB Kafka or vDB OpenSearch cluster, so you can store, search, and analyze logs on your own terms.

Two independent destinations: vDB Kafka (mTLS or SASL authentication) and vDB OpenSearch.
Enable, switch Cluster, or disable export right from the vLB log mapping screen.
Learn more at Working with vLB-Log.

VKS - Usage & Limits

The VKS Portal now includes a Usage & Limits page, giving you visibility into your current resource usage and configuration limits — without having to contact support to check remaining quota.

View the number of Kubernetes Clusters in use vs. the account limit (with a percentage bar).
View configuration limits: Node Groups / Cluster and Nodes / Node Group.
Click Request limit increase to submit a limit increase request to the 24/7 support team.
Click View resource limits on vServer ↗ to view the full vCPU, RAM, and Disk limits on vServer.
Learn more at Usage & Limits.

GreenNode AgentBase — Phase 2

GreenNode AgentBase releases Phase 2 with 7 new features, completing the production-grade AI Agent platform for developer teams:

Marketplace: Browse and 1-click deploy AI agents from a template library — AI Chat, Coding, Automation categories.
Container Registry: Private container image repository automatically provisioned for the organization; used directly when creating a Runtime.
Rate Limit: Control API call frequency by requests or tokens per model and API key.
MCP Governance: MCP Gateway (centralized proxy + policy enforcement) and Policy Groups (access control by tool name, input, and output patterns).
AI Coding: Connect Claude Code and OpenAI-compatible tools to GreenNode MaaS, billed via internal credit-tokens.
Usage & Budget: Real-time cost dashboard by agent/model/API key; set monthly budget limits with automatic alerts at 80% and 100%.
Private Networking: VPC Peering for Agent Runtime and MCP Gateway — agents reach internal services without internet exposure.
Learn more at AI Stack — Release Notes.

April 2026

Agentbase — OpenClaw 1-Click

GreenNode has released OpenClaw 1-Click on Agentbase, enabling anyone to deploy a personal AI Agent powered by OpenClaw directly from the Agent Marketplace — no technical knowledge required, no manual setup, ready in just 40–60 seconds.

Auto-connected to GreenNode MaaS: GreenNode users are automatically granted LLM access without manually configuring API keys. Default model: qwen3-5-27b.
BYOK — Bring Your Own Key: Supports users who want to bring their own API key from external providers (OpenAI, Anthropic, Gemini...).
Channel integrations: Configure Telegram and Zalo connections directly during the deploy step.
My Agents: Manage all OpenClaw instances with status filtering, stop, restart, and delete.
Learn more at OpenClaw 1-Click.

vDB - PostgreSQL Cluster (RDS)

GreenNode has released the PostgreSQL Cluster feature for the vDB Relational service, enabling PostgreSQL deployment with a High Availability architecture, automatic Failover, and flexible scaling for production workloads.

High Availability architecture: Supports 2 to 10 nodes with automatic failover when the Writer encounters an issue.
Two separate RW & RO Endpoints: easily scale out heavy-read workloads.
Built-in vBackup integration: easily configure Backup Policies and Vault Lock in Backup Center.
Popular extensions supported: pgvector, timescaledb, postgis, pg_stat_statements,... suitable for use cases like VectorDB for AI, high-performance real-time analytics, ...
Learn more at PostgreSQL Cluster.

March 2026

Agentbase — AI Agent Infrastructure Platform

Runtime Service: Deploy agents as containers with autoscaling, versioning, and zero-downtime deployment.
Identity Service: Manage agent identity and automatically inject credentials into containers at runtime.
Memory Service: Store conversation history (short-term) and extract semantic facts (long-term).
GreenNode MaaS integration: Direct connection to LLM models via an OpenAI-compatible API.
Learn more at Agentbase.

VKS - Multi-AZ Control Plane

VKS (VNG Kubernetes Engine) now officially supports deploying Kubernetes Clusters with the Control Plane distributed across multiple Availability Zones (AZs), ensuring High Availability for your cluster. If one AZ experiences an outage, the Control Plane continues to operate normally thanks to redundant instances in other AZs. This feature can be combined with both Public Cluster and Private Cluster.

Free of charge during the initial release period. Official pricing will be announced at a later date.
Learn about the concept and management at Multi-AZ Control Plane.
Step-by-step creation guide at Create a Multi-AZ Cluster.
API documentation: VKS API | Terraform documentation: VKS Terraform.

March 2026

Agentbase

GreenNode has released Agentbase — a purpose-built infrastructure platform for AI Agents, enabling developers to deploy and operate AI Agents on the cloud without managing servers, scaling, or credentials. Agentbase includes built-in Identity Service, Runtime Service, and Memory Service, with direct integration to GreenNode MaaS (OpenAI-compatible). Learn more at Agentbase.

January 2026

Model as a Service (MAAS)

MAAS provides AI models as a service, making it easy for customers to use models on a centralized management platform in a simple and cost-effective way.

Learn more about how to create the service here or experience the service now here.

vWAF (Web Application Firewall)

vWAF is an Nginx-based web application firewall service that acts as a reverse proxy to filter and monitor HTTP/S traffic between users and web applications. In terms of application protection features, vWAF supports IP Geolocation, IP Blacklist/Whitelist, NDay Rules - Application Rules, CSP Header, HTTP/2, and OWASPTop 10 protection (including SQLi, XSS, CSRF, RCE, SSRF). vWAF provides bot and DDoS protection, including features such as Anti-bot protection, Advanced Bot Management, Whitelist Good Bots, and prevention of common web attacks like SQLi, XSS, and RCE at the application layer (Layer 7). In addition, vWAF offers management, monitoring, and reporting features such as Reporting, API Access, Realtime Dashboard, Multi-User Management, Request Logs, and Audit Logs.

Learn how to set up the service here or experience the service now here.

Previous2025 NextService package limits

Last updated 6 days ago