Skip to content

User Starter Kit

Welcome to AKKO. This guide walks you through your first steps on the platform -- from logging in to running your first query.

First Login via Keycloak SSO

All AKKO services use single sign-on (SSO) through Keycloak. You only need to log in once.

  1. Open the Cockpit portal in your browser: https://cockpit.<domain>
  2. You will be redirected to the Keycloak login page
  3. Enter your credentials (provided by your administrator)
  4. After authentication, you are redirected back to the Cockpit

One login, all services

Once authenticated through Keycloak, you can access JupyterHub, Superset, Airflow, Grafana, and all other services without logging in again.

The Cockpit is your central hub. It displays all available services as cards with real-time health indicators.

Service Categories

  • Data & Analytics -- Superset (dashboards), Trino (SQL queries), Spark (compute)
  • Notebooks -- JupyterHub (Python, R, Julia notebooks)
  • AI / ML -- Ollama (local LLM), LiteLLM (AI gateway), MLflow (experiment tracking)
  • Governance -- OpenMetadata (data catalog), OPA (access policies)
  • Infrastructure -- MinIO (object storage), PostgreSQL (databases), Keycloak (SSO)
  • Monitoring -- Grafana (dashboards), Prometheus (metrics), Alertmanager (alerts)

Each card shows:

  • Service name and version
  • Health status (green = healthy, red = down)
  • Direct link to the service UI

First Notebook in JupyterHub

  1. Click JupyterHub in the Cockpit
  2. Your personal notebook server starts automatically
  3. You land in JupyterLab with pre-installed kernels: Python 3, R, Julia

Quick Spark Connect Example

Create a new Python notebook and run:

from pyspark.sql import SparkSession

spark = SparkSession.builder \
    .remote("sc://spark-connect:15002") \
    .getOrCreate()

# Read an Iceberg table
df = spark.sql("SELECT * FROM polaris.banking.transactions LIMIT 10")
df.show()

Quick Trino Example

from trino.dbapi import connect

conn = connect(
    host="trino",
    port=8080,
    user="alice",
    catalog="iceberg",
    schema="banking"
)
cursor = conn.cursor()
cursor.execute("SELECT * FROM transactions LIMIT 10")
rows = cursor.fetchall()
for row in rows:
    print(row)

Pre-loaded demo notebooks

Check the shared/ folder in JupyterHub for ready-to-run demos: banking analysis, RAG pipeline, and Spark Iceberg examples.

First SQL Query in Superset SQL Lab

  1. Click Superset in the Cockpit
  2. Navigate to SQL Lab > SQL Editor
  3. Select the Trino database connection (pre-configured)
  4. Choose schema banking from the dropdown
  5. Run your first query:
SELECT
    account_type,
    COUNT(*) AS num_transactions,
    ROUND(AVG(amount), 2) AS avg_amount,
    ROUND(SUM(amount), 2) AS total_amount
FROM iceberg.banking.transactions
GROUP BY account_type
ORDER BY total_amount DESC

Saving Queries

  • Click Save to store your query for later
  • Use Share to send the query link to colleagues

First Dashboard Exploration

  1. In Superset, navigate to Dashboards
  2. Open the AKKO Banking Overview dashboard (pre-provisioned)
  3. Explore the 8 pre-built charts:
    • Transaction volume over time
    • Distribution by account type
    • Geographic distribution
    • Risk score analysis

Interacting with Dashboards

  • Filter -- Use the filter bar at the top to narrow data by date, account type, or region
  • Drill down -- Click chart elements to explore underlying data
  • Export -- Download chart data as CSV or image

All services are accessible via your browser. Replace <domain> with your AKKO deployment domain.

Service URL Purpose
Cockpit https://cockpit.<domain> Central portal
JupyterHub https://jupyter.<domain> Notebooks (Python, R, Julia)
Superset https://superset.<domain> Dashboards & SQL Lab
Airflow https://airflow.<domain> Pipeline orchestration
Grafana https://grafana.<domain> Monitoring dashboards
MinIO Console https://minio.<domain> Object storage browser
Keycloak https://keycloak.<domain> Account management
OpenMetadata https://openmetadata.<domain> Data catalog & governance
MLflow https://mlflow.<domain> ML experiment tracking

Need help?

Contact your platform administrator or check the Troubleshooting page.