Skip to main content
Learn more about Agentic Mode.

​
Administrator-only configuration

The installation and configuration process in this manual should be performed by system administrators only.

​
Requirements

Before proceeding with the installation, ensure the following prerequisites are met:

​
Database

Postgres v17+ must be deployed. Managed cloud solutions like AWS RDS or GCP CloudSQL are recommended. The database should follow a few minimum specifications:
  • 2 vCPU
  • 16 GB of memory
  • 50 GB of disk space
  • Accessible from the EC2 instance running the Qodo stack
If you’re using a public cloud, managed solutions like GCP CloudSQL and AWS RDS are recommended.

​
Qodo stack version

Ensure Qodo Stack version is at least 0.1.300.

​
Helm installation

Ensure Helm is installed for managing Kubernetes deployments. Embedded Cluster (Optional): If Kubernetes is not available, installation can be performed using Replicated (web UI-based deployment).

​
Configuration details for Helm users

​
Environment variables

  • The primary configuration file is /copilot_proxy/settings_prod/.secrets.toml.
  • The file is a Kubernetes secret that must be appended to.

​
Required installations

  • In addition to setting values, you must enable cronjobs and jobs in the Helm chart.
  • The following command ensures all necessary jobs are scheduled: 
kubectl apply -f cronjobs.yaml

​
Installation steps for Helm users

​
1. Database setup

Deploy a Postgres v17+ database and configure it with the required credentials. Example Configuration: 
[agenticdb]
username=<your-username>
password=<your-password>
host=<your-db-host>
port=5432
db_name="agentic"
  • If using a public cloud, managed solutions like AWS RDS, GCP CloudSQL or Azure SQL are recommended.

​
2. Update Helm values

Modify the Helm values file to include the necessary configurations for Agentic Mode: 
jobs:
  db-migration:
    annotations:
      "helm.sh/hook": pre-install,pre-upgrade
      "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded
    command:
      - python
      - -m
      - alembic
      - upgrade
      - head
  db-partition-create-init:
    backoffLimit: 10
    activeDeadlineSeconds: 3600
    annotations:
      helm.sh/hook: post-install,post-upgrade
      helm.sh/hook-weight: "-4"
      helm.sh/hook-delete-policy: hook-succeeded
    command:
      - python
      - /copilot_proxy/db/partitoning/partition_script.py
cronJobs:
  db-partition-delete:
    backoffLimit: 10
    activeDeadlineSeconds: 3600
    concurrencyPolicy: Forbid
    schedule: "0 1 * * *"
    command:
      - /bin/sh
      - -c
      - "PYTHONPATH=/copilot_proxy:/ragger/indexing:/ python /copilot_proxy/db/partitoning/cleanup_script.py"
  db-partition-create:
    concurrencyPolicy: Forbid
    schedule: "0 1 * * *"
    command:
      - python
      - /copilot_proxy/db/partitoning/partition_script.py

​
3. Run Helm upgrade

Apply the updated Helm values by running: 
helm upgrade qodo-gen oci://artifacts-self-hosted.qodo.ai/codium-stack/stable/module -f ./values.yaml
This will run database migrations and create the necessary partitions.

​
4. Update secrets

Modify the Kubernetes secret file /copilot_proxy/settings_prod/.secrets.toml to include: 
[agenticdb] # Agentic DB details
username=$USERNAME
password=$PWD
host=$HOST
port="5432"
db_name="agentic"

[app] # Feature flags
agentic_support = true    # Enable Agentic mode
lean_agent_enabled = true # Enable lean agent for smoother, more flexible development workflows.
custom_mcp_enabled = true # Enable Agentic Tools (MCPs)
apply_flow_on = true
test_flow_on = true

/copilot_proxy/settings/settings.toml (default settings in the image)
[app]
apply_flow_on = false
test_flow_on = false
agentic_support = false
custom_mcp_enabled = false
langgraph_checkpointer="postgres"
Ensure the pods pick up the new version of the secret. You can also pass over these secrets as environemt variables: 
APP.AGENTIC_SUPPORT=true
APP.LEAN_AGENT_ENABLED=true

​
Configuration sections explained

  • [app]: Defines feature flags.
  • [agenticdb]: Contains Agentic DB details.

​
5. Verify Agentic Mode

Restart your IDE and verify that Agentic Mode is enabled.

​
Installation steps for Embedded Cluster users

For customers without Kubernetes, use Replicated to install Agentic Mode on a virtual machine.

​
Steps

1
Enable the Agentic Mode license flag: Notify the Qodo team to enable the license flag.
2
Sync license: Once confirmed, sync the license in the Admin web UI.
3
In the Admin web UI, navigate to the Config tab.
4
Enable Agentic Mode within Qodo Gen: In the Qodo Gen Settings section, enable Agentic Mode.
5
Update secret: Append to the end of Qodo Gen secret toml file:
[agenticdb]
username=<your-username>
password=<your-password>
host=<your-db-host>
port=5432
db_name="agentic"
6
Ensure the configuration was updated: ensure the popup The config for Qodo stack has been updated. appears, then click Go to updated version.
7
Deploy the updated configuration: in the Version history tab you’ve been forwarded to, find the Config change entry in the top of the versions list. Click Deploy.
8
Wait for a pod to start: Changing the secret causes the container restart. Wait until it fully starts up.
9
Restart and start using Agentic Mode: Restart your IDE. You should be able to use Agentic Mode in your IDE.
For further details, refer to the official documentation.