To use ToolJet AI features in your deployment, make sure to whitelist `https://api-gateway.tooljet.com` and `https://python-server.tooljet.com` in your network settings.
You should set up a PostgreSQL database manually to be used by ToolJet. We recommend using an **RDS PostgreSQL database**. You can find the system requirements [here](/docs/3.5.0-LTS/setup/system-requirements#postgresql).
ToolJet runs with **built-in Redis** for multiplayer editing and background jobs. When running **separate worker containers** or **multi-pod setup**, an **external Redis instance** is **required** for job queue coordination.
1. Create an EKS cluster and connect to it to start with the deployment. You can follow the steps as mentioned in the [AWS documentation](https://docs.aws.amazon.com/eks/latest/userguide/create-cluster.html).
The file below is just a template and might not suit production environments. You should download the file and configure parameters such as the replica count and environment variables according to your needs.
Make sure to edit the environment variables in the `deployment.yaml`. We advise using secrets to set up sensitive information. You can check out the available options [here](/docs/setup/env-vars). <br/>
PG_DB=tooljet_production # Must be a unique database name (do not reuse across deployments)
```
Make sure to edit the environment variables in the `deployment.yaml`. You can check out the available options [here](/docs/setup/env-vars).
#### SSL Configuration for AWS RDS PostgreSQL
:::warning
**Important**: When connecting to PostgreSQL 16.9 on AWS RDS with SSL enabled, you need to configure SSL certificates. The `NODE_EXTRA_CA_CERTS` environment variable is critical for resolving SSL certificate chain issues and for connecting to self-signed HTTPS endpoints.
:::
For AWS RDS PostgreSQL connections, create a ConfigMap with the certificate:
3. Create a Kubernetes service to publish the Kubernetes deployment that you have created. We have a [template](https://tooljet-deployments.s3.us-west-1.amazonaws.com/kubernetes/service.yaml) for exposing the ToolJet server as a service using an AWS Load Balancer.
Use the ToolJet-hosted database to build apps faster, and manage your data with ease. You can learn more about this feature [here](/docs/tooljet-db/tooljet-database).
Deploying ToolJet Database is mandatory from ToolJet 3.0 or else the migration might break. Checkout the following docs to know more about new major version, including breaking changes that require you to adjust your applications accordingly:
The **`PGRST_DB_URI`** variable is **required** for PostgREST, which exposes the database as a REST API. This must be explicitly set for proper functionality.
**Ensure these configurations are correctly set up before proceeding with the ToolJet deployment. Make sure these environment variables are set in the same environment as the ToolJet deployment.**
ToolJet Workflows allows users to design and execute complex, data-centric automations using a visual, node-based interface. This feature enhances ToolJet's functionality beyond building secure internal tools, enabling developers to automate complex business processes.
- **WORKER** (required): Enables job processing. Set to `true` to activate workflow scheduling
- **TOOLJET_WORKFLOW_CONCURRENCY** (optional): Controls the number of workflow jobs processed concurrently per worker instance. Default is 5 if not specified
:::warning
**External Redis Requirement**: When running separate worker containers or multiple instances, an external stateful Redis instance is **required** for job queue coordination. The built-in Redis only works when the server and worker are in the same container instance (single instance deployment).
:::
#### Deploying Redis for Workflows
Deploy a stateful Redis instance using the following example configuration:
<summary>Built-in Redis vs External Redis</summary>
ToolJet images include a built-in Redis instance for development. When deploying workflows in production, you must update your deployment configuration to use the external stateful Redis:
Change **REDIS_HOST** from **localhost** to **redis-service** in your deployment YAML:
- ConfigMap with production-ready Redis configuration
- A Secret for optional password authentication
:::info
This is an example configuration that you can customize to your needs. However, **AOF (Append Only File) persistence** and **`maxmemory-policy noeviction`** are critical settings that must be maintained for BullMQ job queue reliability.
:::
After deploying Redis, configure ToolJet to connect to it using these environment variables in your deployment:
**Note:** Ensure that these environment variables are added to your Kubernetes deployment configuration (e.g., in your deployment.yaml file or Kubernetes secret).
New LTS versions are released every 3-5 months with an end-of-life of at least 18 months. To check the latest LTS version, visit the [ToolJet Docker Hub](https://hub.docker.com/r/tooljet/tooljet/tags) page. The LTS tags follow a naming convention with the prefix `LTS-` followed by the version number, for example `tooljet/tooljet:ee-lts-latest`.
If this is a new installation of the application, you may start directly with the latest version. This guide is not required for new installations.
#### Prerequisites for Upgrading to the Latest LTS Version:
- It is crucial to perform a **comprehensive backup of your database** before starting the upgrade process to prevent data loss.
- Users on versions earlier than **v2.23.0-ee2.10.2** must first upgrade to this version before proceeding to the LTS version.
_If you have any questions feel free to join our [Slack Community](https://join.slack.com/t/tooljet/shared_invite/zt-2rk4w42t0-ZV_KJcWU9VL1BBEjnSHLCA) or send us an email at support@tooljet.com._