# 2025.38 Infrastructure Changes

1 min read

# Changes on Application environment

| **Change** | **Name** | **Default** | **Example** | **Required** | **Applications** | **Short Description** |
| Removed | `FEATURE_FLAG_AGENTIC_TABLE_ONBOARDING_V2_UN_12041` | `true` | `true` | No | `next-chat` | Enables single-page sheet creation UI for the Agentic Table. <br>Can be removed since it is GA now. |

* * *

## Unique AI Agent

To be able to use Unique AI, update the `INTERNAL_MODULES_CONFIG` environment variable in the `node-chat` application

Add the following line

```
{"name": "UniqueAi", "apiURL": "http://assistants-core.chat.svc.cluster.local:8081/core/unique_ai"}
```

_Note: the internal module config has to be a valid json_

* * *

**API Request Timeout**

We discovered that newer thinking models tend to take a long time to respond. Especially for chat-completion interactions and not streaming.

To increase the request timeout Kong Application gateway, the following change needs to be done:

- Chat backend application YAML (node-chat):

```
...
service:
  ...
  annotations:
    ...
    konghq.com/read-timeout: "600000" # 10 minutes
    konghq.com/write-timeout: "600000" # 10 minutes
    konghq.com/connect-timeout: "600000" # 10 minutes
```

This will increase the timeout of all requests to the backend chat application (used for the public API and LXM interactions) to 10 Minutes.

In order to be able to set these annotations, you have to use backend-service helm chart in version >=5.5.0

| **Author** | Solution Engineering |