<!-- Body -->
<!-- Callout -->
<!-- Section 1 -->
1. What You Get Free
Azure Databricks offers two free tiers. For this masterclass, you want the 14-Day Premium Trial — not the personal “Free Edition”. Here’s the difference:
<table class="w-full text-sm border-collapse">
<thead>
<tr class="border-b border-border-color">
<th class="text-left py-2 pr-4 text-text-muted font-600 text-xs uppercase tracking-wider">Feature</th>
<th class="text-left py-2 pr-4 text-text-muted font-600 text-xs uppercase tracking-wider">14-Day Premium Trial ✅</th>
<th class="text-left py-2 text-text-muted font-600 text-xs uppercase tracking-wider">Free Edition ❌</th>
</tr>
</thead>
<tbody class="divide-y divide-border-color/50">
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Spark clusters</td>
<td class="py-3 pr-4 text-text-muted">Full multi-node + single-node</td>
<td class="py-3 text-text-muted">Serverless only, quota-limited</td>
</tr>
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Delta Lake & Unity Catalog</td>
<td class="py-3 pr-4 text-text-muted">Fully included</td>
<td class="py-3 text-text-muted">Limited</td>
</tr>
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Job orchestration</td>
<td class="py-3 pr-4 text-text-muted">Full Workflows engine</td>
<td class="py-3 text-text-muted">Not available</td>
</tr>
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Cloud requirement</td>
<td class="py-3 pr-4 text-text-muted">Azure subscription needed</td>
<td class="py-3 text-text-muted">No cloud account needed</td>
</tr>
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Duration</td>
<td class="py-3 pr-4 text-copper font-600">14 days (free DBUs)</td>
<td class="py-3 text-text-muted">No expiry, but daily quota capped</td>
</tr>
</tbody>
</table>
<!-- Section 2 -->
2. Prerequisites
You need these before starting — workspace creation takes under 5 minutes once you have them:
✓ An active Azure subscription — Pay-As-You-Go or an existing Azure Free Account. The Databricks 14-day trial does not work on an “Azure Free Trial” subscription ($200 credit trial). If you are on that trial, first upgrade to Pay-As-You-Go under Subscriptions → Upgrade in the Azure Portal.
✓ A resource group — you can create one during setup. Name it
rg-databricks-masterclass.✓ Sufficient vCPU quota — most new subscriptions have enough (4–8 vCPUs). If creation fails with a quota error, see troubleshooting at the bottom.
<!-- Section 3 -->
3. Create Your Databricks Workspace
1
Open the Azure Portal and search for Databricks
Go to portal.azure.com. In the top search bar, type Azure Databricks and click it under “Services”. Then click + Create.
2
Fill in the Basics tab
Subscription: select your Pay-As-You-Go subscription
Resource group: create new →
rg-databricks-masterclassWorkspace name: e.g.
datakern-masterclassRegion: West Europe or North Europe (low latency from EU)
3
⚡ Select the Trial pricing tier — this is the critical step
In the Pricing tier dropdown, select Trial (Premium - 14-Days Free DBUs). This is what unlocks 14 days of free compute. If you select “Premium” without “Trial”, you will be billed immediately. Note: the Standard tier was retired in April 2026 and is no longer available for new workspaces.
4
Leave Networking as default
For the masterclass, default network settings are fine. Do not enable VNet injection — it adds complexity and cost you don’t need.
5
Review + Create
Click Review + create. Confirm the summary shows “Trial (Premium)” in the pricing tier, then click Create. Deployment takes 2–4 minutes.
6
Launch the workspace
Click Go to resource, then Launch Workspace. The Databricks UI opens in a new tab — your workspace is live.
<!-- Section 4 -->
4. Create Your First Cluster (Cost-Safe)
Inside your Databricks workspace, follow these exact settings to keep costs minimal during the masterclass:
1
Go to Compute → Create compute
In the left sidebar, click Compute, then click Create compute.
2
Name it and set Single Node mode
Name it masterclass-cluster. Under Cluster mode, select Single Node. This keeps everything on one VM — more than enough for the masterclass, and significantly cheaper than multi-node.
3
⚡ Set Auto Termination to 15 minutes
Under Auto Termination, tick the checkbox and enter 15 minutes. A Databricks cluster left running overnight can consume €10–30 in infrastructure charges. This one setting prevents that entirely.
4
Choose Standard_DS3_v2 as the node type
Under Node type, pick Standard_DS3_v2 (14 GB RAM, 4 vCPUs). This is the recommended balance for the masterclass — enough memory for real data workloads without the cost of larger instances.
5
Select Databricks Runtime LTS version
Select the latest LTS (Long Term Support) runtime. LTS releases are stable and well-tested — avoid preview or ML runtimes unless the instructor specifies otherwise.
6
Click Create compute
Click Create compute. The cluster starts in 3–5 minutes and shows a green dot when running. You’re ready.
<!-- Section 5 -->
5. Avoid Unexpected Charges
→ Terminate the cluster when done. Go to Compute in the sidebar and click Terminate after each session. Auto-termination is your backup, not your primary safeguard.
→ After the masterclass, delete the resource group. In the Azure Portal: Resource groups → rg-databricks-masterclass → Delete. This removes the workspace, managed storage, and all underlying resources in one click.
→ Set a cost alert in Azure. Go to Cost Management → Budgets → + Add. Set a €5 budget with email alerts at 50% and 100%. You’ll be notified before any real spending builds up.
→ After 14 days, the workspace switches to paid Premium. If you don’t delete the workspace within 14 days of creation, it automatically converts to standard Premium billing. Set a calendar reminder on the day you create it.
<!-- Section 6: Cost table -->
6. Estimated Session Cost
A typical 3-hour masterclass session on a Single Node cluster costs approximately:
<table class="w-full text-sm border-collapse">
<thead>
<tr class="border-b border-border-color">
<th class="text-left py-2 pr-4 text-text-muted font-600 text-xs uppercase tracking-wider">Resource</th>
<th class="text-left py-2 pr-4 text-text-muted font-600 text-xs uppercase tracking-wider">Usage (3h session)</th>
<th class="text-left py-2 text-text-muted font-600 text-xs uppercase tracking-wider">Est. Cost</th>
</tr>
</thead>
<tbody class="divide-y divide-border-color/50">
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Databricks DBUs (Premium)</td>
<td class="py-3 pr-4 text-text-muted">~6 DBUs (2 DBU/hr × 3h)</td>
<td class="py-3 text-copper font-600">Free (trial covers this)</td>
</tr>
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Azure VM (DS3_v2)</td>
<td class="py-3 pr-4 text-text-muted">3 hours compute</td>
<td class="py-3 text-text-muted">~€0.60</td>
</tr>
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Managed Disk (OS)</td>
<td class="py-3 pr-4 text-text-muted">30 GB standard SSD</td>
<td class="py-3 text-text-muted">~€0.02/day</td>
</tr>
<tr>
<td class="py-3 pr-4 font-600 text-text-main">ADLS Gen2 Storage</td>
<td class="py-3 pr-4 text-text-muted">Minimal (dataset files)</td>
<td class="py-3 text-copper font-600">Covered by Azure free tier</td>
</tr>
<tr>
<td class="py-3 pr-4 font-600 text-text-main">Total (3h session)</td>
<td class="py-3 pr-4 text-text-muted">Full cluster, active work</td>
<td class="py-3 text-green-500 font-700">≈ €0.62</td>
</tr>
</tbody>
</table>
<!-- Section 7: Troubleshooting -->
7. Common Issues & Fixes
<!-- Section 8: Summary -->
8. Summary
You now have a fully functional Azure Databricks Premium workspace with a cost-safe Single Node cluster ready for the masterclass. The entire setup takes under 10 minutes. The free trial covers all Databricks charges for 14 days — your only exposure is minimal Azure VM cost (~€0.60 for the 3-hour session), which is covered if you have the Azure $200 free credit.
Create this workspace at least the day before the masterclass. Cluster startup takes 3–5 minutes — you don’t want to wait for it while the session is already running.
<!-- References -->