> For the complete documentation index, see [llms.txt](https://docs.ionos.com/cloud/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ionos.com/cloud/release-notes/october-2026.md).

# October 2026

IONOS CLOUD release notes: new features, improvements, fixes, and deprecations for October 2026.

<a href="https://release-notes-feed.s3.eu-central-3.ionoscloud.com/feed.rss" class="button primary" data-icon="rss">Subscribe through RSS feed</a>

## What's new

### [<mark style="color:blue;">**Predefined Grafana dashboards for DBaaS PostgreSQL, MariaDB, and In-Memory DB**</mark>](/cloud/databases/databases/dbaas-grafana-dashboards.md)

{% hint style="success" %}
**October 19**

Your Grafana instance for the <code class="expression">space.vars.ionos\_cloud</code> **Monitoring Service** and **Logging Service** now includes six predefined dashboards for **DBaaS PostgreSQL**, **DBaaS MariaDB**, and **DBaaS In-Memory DB**. Each database has one metrics dashboard and one logs dashboard. To open them, go to **Dashboards** > **IONOS CLOUD** in Grafana.

The dashboards populate automatically with the metrics and logs of every cluster that forwards them. To forward metrics and logs, activate [<mark style="color:blue;">Central Monitoring</mark>](/cloud/observability/monitoring-service/dcd-how-tos/central-monitoring.md#turn-on-central-monitoring-for-a-region) and [<mark style="color:blue;">Central Logging</mark>](/cloud/observability/logging-service/dcd-how-tos/central-logging.md#turn-on-central-logging-for-a-region) for the region of the cluster, and turn on **Enable metrics** and **Enable logs** in the **Observability** settings of the cluster. For the content of each dashboard, see [<mark style="color:blue;">Predefined DBaaS Grafana Dashboards</mark>](/cloud/databases/databases/dbaas-grafana-dashboards.md).

* **Metrics dashboards:** A [<mark style="color:blue;">fleet table</mark>](/cloud/databases/databases/dbaas-grafana-dashboards.md#dashboard-coverage) lists every cluster that sends metrics. Select a cluster to see its health, connections, throughput, replication, storage usage, and instance resource usage.
* **Logs dashboards:** A fleet table lists every cluster that sends logs. Select a cluster to see its errors, warnings, slow queries, deadlocks, and failed logins, or snapshot, replication, and client connection events for In-Memory DB. You can also search the full cluster log.

The metrics and logs dashboards of each database link to each other and keep the selected cluster and time range.
{% endhint %}

### [<mark style="color:blue;">**OpenAI-compatible rate limit headers for AI Model Hub**</mark>](/cloud/ai/ai-model-hub/how-tos/rate-limits.md)

{% hint style="success" %}
**October 12**

The AI Model Hub rate limit headers now follow the OpenAI convention. `X-RateLimit-Remaining-Requests` counts down from the base limit and leaves out the burst allowance. At the default limit of 5 requests per second, the first request now returns `4` instead of `9`. The new `X-RateLimit-Reset-Requests` header gives the time until the limit is back at its initial state, as a duration such as `200ms`, `1.5s`, or `6m0s`. A `429 Too Many Requests` response now includes a `Retry-After` header with the number of seconds to wait before retrying. Client libraries that honor `Retry-After`, such as the `openai` Python library, handle this for you. For more information, see [<mark style="color:blue;">Rate Limits</mark>](/cloud/ai/ai-model-hub/how-tos/rate-limits.md#hitting-rate-limits).
{% endhint %}

### [<mark style="color:blue;">**Cloud GPU Virtual Machines available in region Worcester (UK)**</mark>](/cloud/compute-services/compute-engine/cloud-gpu-vm.md)

{% hint style="success" %}
**October 7**

GPU Cloud Virtual Machines are now available in the region **Worcester (UK)** `gb/bhx`, expanding the <code class="expression">space.vars.ionos\_cloud</code> GPU-accelerated compute capabilities to a new geographic location. You can now deploy and manage GPU instances directly through the Data Center Designer or CloudAPI V6 and SDKs or Configuration Management Tools.

GPU VMs in both locations, **Frankfurt** and **Worcester** are available to all <code class="expression">space.vars.ionos\_cloud</code> customers starting today.
{% endhint %}

### [<mark style="color:blue;">**Advisory on Managed Kubernetes Vulnerability**</mark>](/cloud/security-safeguards/vulnerability-register/2026/managed-kubernetes-vulnerability.md)

{% hint style="success" %}
**October 6**

Information on the security advisory for **Managed Kubernetes Worker Node Credential Exposure** is now available in the documentation portal. Disclosed on July 16, 2026, this vulnerability affects [<mark style="color:blue;">Managed Kubernetes</mark>](/cloud/containers/managed-kubernetes.md) and caused worker nodes running Managed Kubernetes v1.33+ to expose an internal administrative credential via the cloud-init provisioning mechanism.

<code class="expression">space.vars.ionos\_cloud</code> has applied updated Managed Kubernetes node images across all regions to mitigate the vulnerability.
{% endhint %}

## Documentation updates

### [<mark style="color:blue;">**Monitor OpenClaw Usage and Cost with the Monitoring Service**</mark>](https://docs.ionos.com/cloud/tutorials/observability/monitoring-service/monitor-openclaw-with-monitoring-service)

{% hint style="success" %}
**October 13**

A new tutorial shows how to export OpenClaw gateway telemetry to <code class="expression">space.vars.ionos\_cloud</code> observability services through the community `otel-observability` plugin. The plugin configures each signal independently, so you can send metrics to the **Monitoring Service**, logs to the **Logging Service**, and traces to the **Tracing Service**, in any combination.

The tutorial also provides a Grafana dashboard for the metrics, reporting token usage, implied cost, request latency, tool call activity, and security guardrail signals. Panels break these down by model and by conversation.
{% endhint %}

### [<mark style="color:blue;">**Build a Hardened Kernel for IONOS CLOUD Confidential VMs**</mark>](/cloud/compute-services/compute-engine/confidential-computing-vm/how-tos/hardened-cvm-kernel-howto.md)

{% hint style="success" %}
**October 8**

A new how-to guide covers how to compile a minimal Linux kernel for Confidential VMs that removes console, input, serial, USB, and management-channel drivers at compile time, eliminating the code paths a hypervisor could otherwise use to interact with the guest. The guide includes kernel configuration using the Debian cloud baseline, package building, GRUB hardening, and post-restart verification over SSH. Validated on Debian 13 (Trixie) with kernel 6.12.101.
{% endhint %}

### [<mark style="color:blue;">**Red Hat OpenShift 4.20 deployment guide**</mark>](/cloud/containers/red-hat-openshift.md)

{% hint style="success" %}
**October 5**

The [<mark style="color:blue;">Red Hat OpenShift</mark>](/cloud/containers/red-hat-openshift.md#set-up-red-hat-openshift-on-ionos-cloud) deployment guide has been updated to version 4.20. Red Hat OpenShift Container Platform 4.20 is now officially validated on <code class="expression">space.vars.ionos\_cloud</code> infrastructure as a Certified Cloud Service Provider (CCSP). Red Hat CCSP partners can download the updated deployment guide for step-by-step instructions on deploying and configuring a Red Hat OpenShift Container Platform 4.20 cluster on <code class="expression">space.vars.ionos\_cloud</code>.
{% endhint %}

## Deprecations

### [<mark style="color:blue;">**Windows Server 2016 Public Image deprecation**</mark>](/cloud/backup-and-storage/images-snapshots/public-images.md)

{% hint style="success" %}
**October 5**

Effective **January 12, 2027**, Microsoft ends extended support for **Windows Server 2016**. After this date, it will no longer be offered as a public image for new deployments. While existing VMs running Windows Server 2016 remain active, they will stop receiving security patches. We recommend migrating to Windows Server 2019, 2022, or 2025 before the deadline.
{% endhint %}

### [<mark style="color:blue;">**Cloud API v6 user password becomes read-only**</mark>](https://api.ionos.com/docs/cloud/v6/#tag/User-management)

{% hint style="success" %}
**October 8**

Effective **November 5, 2026**, the `password` property of the Cloud API v6 user management endpoint `PUT /um/users/{userId}` becomes read-only for existing users. You can set a password only when you create a user with `POST /um/users`. Contract owners and administrators can no longer change the password of another user through the API. Users change their own password in the <code class="expression">space.vars.ionos\_cloud</code> DCD under **Menu** > **Your Profile** > **Account Security**, as described in [<mark style="color:blue;">Account Security</mark>](/cloud/set-up-ionos-cloud/account-management/account-security.md), or [<mark style="color:blue;">reset a forgotten password</mark>](https://www.ionos.com/help/my-account/access-passwords/resetting-your-ionos-account-password/). From this date, a `PUT /um/users/{userId}` request that includes `password` returns an error stating that the field is immutable and must be removed from the payload, and the request is not executed. Remove the `password` property from any `PUT /um/users/{userId}` requests before the effective date.
{% endhint %}

### [<mark style="color:blue;">**FLUX.1-schnell and Paraphrase Multilingual MPNet v2 scheduled for retirement**</mark>](/cloud/ai/ai-model-hub/models.md)

{% hint style="success" %}
**October 21**

**FLUX.1-schnell** and **Paraphrase Multilingual MPNet v2** are scheduled to be retired from the AI Model Hub on **November 30, 2026**. Migrate FLUX.1-schnell workloads to [<mark style="color:blue;">FLUX.2-klein-4B</mark>](/cloud/ai/ai-model-hub/models/image-generation-models/flux-2-klein-4b.md), and Paraphrase Multilingual MPNet v2 workloads to [<mark style="color:blue;">BGE M3</mark>](/cloud/ai/ai-model-hub/models/embedding-models/bge-m3.md). FLUX.2-klein-4B is charged per megapixel of output plus input images instead of per image, so the cost per request changes. Embeddings from Paraphrase Multilingual MPNet v2 (768 dimensions) and BGE M3 (1024 dimensions) are not interchangeable, so re-embed your existing content with BGE M3 before you switch.
{% endhint %}

### [<mark style="color:blue;">**Llama 3.1 8B End of Life**</mark>](/cloud/ai/ai-model-hub/models.md)

{% hint style="success" %}
**October 21**

Llama 3.1 8B has reached end of life and is no longer available in the AI Model Hub. Requests using the `meta-llama/Meta-Llama-3.1-8B-Instruct` model ID now receive an `HTTP 404 Not Found` response with the `MODEL_NOT_SUPPORTED` error code. Its designated successor is [<mark style="color:blue;">Qwen3.5-9B</mark>](/cloud/ai/ai-model-hub/models/llms/qwen3-5-9b.md).
{% endhint %}

### [<mark style="color:blue;">**Mistral Nemo (12B) and BGE Large 1.5 End of Life**</mark>](/cloud/ai/ai-model-hub/models.md)

{% hint style="success" %}
**October 22**

Mistral Nemo (12B) and BGE Large 1.5 have reached end of life and are no longer available in the AI Model Hub. Requests using the `mistralai/Mistral-Nemo-Instruct-2407` or `BAAI/bge-large-en-v1.5` model ID now receive an `HTTP 404 Not Found` response with the `MODEL_NOT_SUPPORTED` error code. Their designated successors are [<mark style="color:blue;">Qwen3.5-9B</mark>](/cloud/ai/ai-model-hub/models/llms/qwen3-5-9b.md) and [<mark style="color:blue;">BGE M3</mark>](/cloud/ai/ai-model-hub/models/embedding-models/bge-m3.md). Embeddings from BGE Large 1.5 and BGE M3 are not interchangeable, so re-embed your existing content with BGE M3.
{% endhint %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.ionos.com/cloud/release-notes/october-2026.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.