Skip to content

InstallTeadalTools.md

Latest commit

 

History

History
211 lines (134 loc) · 5.76 KB

InstallTeadalTools.md

File metadata and controls

211 lines (134 loc) · 5.76 KB

Teadal tools installation

This document provides a guide to install the tools developed in the TEADAL project. These tools requires a TEADAL node up and running (see Teadal Node installation guide)

At this time, the following tools are available:

If any dependencies are indicated, then it is required to configure the ArgoCD project in your Teadal Node to add the required tool.

Advocate

Dependencies

  • Jaeger

See the related page to know how to install the dependencies

Preliminary steps

Before Advocate will work you will need to create the related namespace

kubectl create namespace trust-plane

Then it is required to configure all needed secrets, variables for Advocate blockchain such as wallet private key, VM key and Ethereum Remote Procedure Call (RPC) Address. For that run this command:

node.config --microk8s advocate

Now you can enter the required values. For the question about the "ADVOCATE_ETH_POA" , enter "1" as value.

Tool deployment

Be sure that, on the repo the kustomization file used by argocd has the line - advocate uncommented. E.g.:

apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization

resources:
- project.yaml
- advocate

Commit the changes to the repo, thus ArgoCD can fetch the updated configuration. Here an example of the possible list of command needed

git add deployment/mesh-infra/argocd/projects/plat-infra-services/
git commit -m "enable advocate"
git push

After few minutes, ArgoCD will realizes the update and starts deploying the related pods.

Checking installation

Check pods that are in Trust-plane namespace:

kubectl get pods -n trust-plane

screenshot

Check the Advocate pod log to make sure that it is up and running:

kubectl logs <advocate-pod-name> -n trust-plane

screenshot

Catalogue deployment

The deployment of the TEADAL data catalogue requires an approval step. Who is interested in this tool must send an e-mail to XXXX@cefriel.com. In case of acceptance, gitlab credentials will be provided. These credentials are required to properly setup the environment for the deployment

Dependencies

Keycloak and Postgres

Preliminary steps

Be sure that the VM on which the catalogue will be installed has a public IP, or it can be reached from the Web.

First step is to create the secrets to allow microk8s to pull the images from the repository

kubectl create secret docker-registry teadal-registry-secret -n catalogue \
  --docker-server=https://registry.teadal.ubiwhere.com \
  --docker-username=<gitlab username> \
  --docker-password="<gitlab password>" \
  --docker-email=<gitlab account email>

In case the microk8s notifies that the namespace catalogue does not exist, the following command solves the issue:

kubectl create namespace catalogue

Tool deployment

Be sure that, on the repo the kustomization file used by argocd has the line - catalogue uncommented. E.g.:

apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization

resources:
- project.yaml
- catalogue

Commit the changes to the repo, thus ArgoCD can fetch the updated configuration. Here an example of the possible list of command needed

git add deployment/mesh-infra/argocd/projects/plat-app-services/
git commit -m "enable catalogue"
git push

After few minutes, ArgoCD will realizes the update and starts deploying the related pods.

Checking installation

In a browser, open the page http://<host>/catalogue. A login page to access to the catalog will appear if everything is properly working.

Policy manager

Dependencies

TBD

Preliminary steps

TBD

Tool deployment

TBD

Checking installation

TBD

AI Driven Performance Monitoring

The AI-DPM tool is composed of two distinct but interdependent components: • AI_api: the backend service responsible for fetching and processing telemetry data using ML models • AI_dashboard: the frontend dashboard used to visualize predictions and interact with the models

Both components must be installed together to ensure full functionality.

Dependencies

To work correctly, the following dependencies must also be enabled and running: • Prometheus (for metrics ingestion) • Thanos (for long-term metrics storage and query) • Kepler (for power-related telemetry collection)

Tool deployment

Be sure that, in the repo, the Istio Kustomization file includes the following resources uncommented:

resources:
  - ai-dashboard
  - ai-prediction-api
  - prometheus
  - kepler
  - thanos

Once the above lines are enabled, apply the configuration by running the following command:

kustomize build deployment/mesh-infra/istio | kubectl apply -f -

This will deploy the AI components as well as the required telemetry infrastructure into the istio-system namespace.

Checking installation

You can verify that all pods are running with:

kubectl get pods -n istio-system | grep ai-

To access the AI dashboard:

kubectl port-forward svc/ai-dashboard 8501:8501 -n istio-system

Then open your browser at http://localhost:8501

To test the prediction API (optional, for debugging):

kubectl port-forward svc/ai-prediction-api 9000:9000 -n istio-system
curl http://localhost:9000/fetch