Modificare numero di core databricks cluster

Question

Modificare numero di core databricks cluster

Cristiano Mombello 10

non riesco a modificare il mio cluster databricks per aggiungere 4 cores. se vado nel clusrter non mostra da nessuna parte il numero di core il che mi pare assurdo, inoltre non mostra la possibilità di modificare la configurazione.

Sono proprietario del cluster ma se vado in quote mi dice che non sono nemmeno Collaboratore.

Ho fatto di tutto.

Gentilmente mi sapete dire cosa devo fare?

0 comments

Answer accepted by question author

Peram Venkateswara Reddy 5 Microsoft External Staff Moderator

Hi @Cristiano Mombello , Hope you are doing well.

Databricks doesn't let you set an exact core count directly. Cluster capacity comes from node type (vCPUs per node) × number of workers, so you change those two things instead. Databricks runs one executor per worker node.

As for why the settings aren't visible — check these in order, cheapest first:

Cluster state

You can only edit a running or terminated cluster. If it's PENDING, RESTARTING, or TERMINATING, the Edit controls are unavailable until it settles. Wait for it to reach Running or Terminated and check again.

Your permission level

There are four levels: NO PERMISSIONS, CAN ATTACH TO, CAN RESTART, and CAN MANAGE. You need CAN MANAGE to change configuration. CAN RESTART lets you start and stop it but not resize it — which matches the symptom of seeing the cluster but not being able to change it.

If you're not sure, ask a workspace admin to check Compute → (your cluster) → kebab menu → Edit permissions. Requesting "CAN MANAGE" by name will get you a much faster answer than "I can't edit it."

The cluster policy

Per the docs: "If you don't see a particular setting in your UI, it's because the policy you've selected does not allow you to configure that setting."

Open the cluster's Configuration tab — the policy name is shown there. Open that policy and you can see exactly which attributes are fixed (locked) versus allowlist/range (adjustable within limits). If node_type_id or num_workers is fixed, a workspace admin has to change the policy or assign you a different one.

What you can do right now

Clone it. Kebab menu → Clone opens the creation UI pre-populated with the same config. If your policy allows creation, you can pick a larger node type or more workers there. Note that permissions and attached notebooks are not carried over.
Use the CLI. These respect policy and ACLs but avoid UI-level gating:
```
  # Change worker count only
```

databricks clusters resize <cluster-id> --num-workers 8

Change node type (requires full config in the JSON)

databricks clusters edit --json @cluster-config.json

  
   You can grab the current config via the JSON button at the top right of the Configuration tab (read-only view) as a starting point.
  
One thing to know before you resize

Changing the number of workers on a running cluster does *not* require a restart. Changing the node type does — and that will interrupt anyone currently attached. Plan accordingly if it's a shared cluster.

On quota: worth separating the two symptoms. If the controls are *missing*, it's state, permissions, or policy. If the controls are *visible but the change fails on apply* with a quota error, then check VM family vCPU quota for that region in the Azure portal. Based on your description it's the former.

References

- [Compute configuration reference](https://learn.microsoft.com/azure/databricks/compute/configure"https://learn.microsoft.com/azure/databricks/compute/configure")

- [Manage classic compute — edit, clone, permissions](https://learn.microsoft.com/azure/databricks/compute/clusters-manage"https://learn.microsoft.com/azure/databricks/compute/clusters-manage")

- [clusters command group (CLI)](https://learn.microsoft.com/azure/databricks/dev-tools/cli/reference/clusters-commands"https://learn.microsoft.com/azure/databricks/dev-tools/cli/reference/clusters-commands")

- [Classic compute configuration best practices — sizing](https://learn.microsoft.com/azure/databricks/compute/cluster-config-best-practices"https://learn.microsoft.com/azure/databricks/compute/cluster-config-best-practices")

Kind Regards,   
Microsoft Support Team.

Cristiano Mombello 10 Reputation points

2026-08-08T06:28:18.5833333+00:00

Thank you @Peram Venkateswara Reddy for the detailed and thorough explanation — I really appreciate the clarity of your guidance.

2 additional answers

Your answer

Cristiano Mombello 10 Reputation points

2026-08-08T06:28:18.5833333+00:00

Thank you @Peram Venkateswara Reddy for the detailed and thorough explanation — I really appreciate the clarity of your guidance.

Answer 1

Deleted

This answer has been deleted due to a violation of our Code of Conduct. The answer was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.

Comments have been turned off. Learn more

Answer 2

Ciao Cristiano,
quello che descrivi di solito succede quando non hai i permessi corretti sul workspace, anche se risulti “proprietario” del cluster. In Databricks, per modificare la configurazione (incluso il numero di core), servono permessi di livello superiore rispetto al semplice “owner”.

Ecco i punti da controllare:

1. Devi avere il ruolo “Can Manage” o essere nel gruppo con permessi di amministrazione Essere proprietario del cluster non basta se il tuo workspace o la tua subscription ha delle quote o dei limiti gestiti dagli admin.

2. La quota dei core è gestita a livello di workspace / cloud provider Se in “Quota” ti dice che non sei Collaboratore, significa che non hai i permessi per aumentare le risorse. Solo un amministratore del workspace può farlo.

3. I cluster non mostrano il numero di core se usi un tipo di cluster con configurazione automatica Alcuni cluster (ad esempio quelli serverless o quelli con policy aziendali) nascondono i dettagli hardware e non permettono modifiche manuali.

4. Potrebbe esserci una Cluster Policy che blocca le modifiche Se la tua azienda ha impostato una policy, non potrai cambiare né core né configurazione finché un admin non la aggiorna.

Cosa devi fare concretamente

Chiedere a un Workspace Admin di aggiungerti come Collaborator o Can Manage.

Far verificare agli admin se c’è una Cluster Policy che blocca le modifiche.

Far controllare la quota dei core nel cloud (Azure/GCP/AWS), perché potresti essere limitato lì.

Finché non hai quei permessi, Databricks non ti farà modificare nulla, anche se sei “owner”.

Cristiano Mombello 10 Reputation points

2026-08-05T09:07:16.3766667+00:00

Dear M. Chester,

thank you very much for your time and for your helpful reply — I really appreciate it.

Just to clarify, my question is not about Databricks workspace permissions or cluster settings inside Databricks. I am referring specifically to Azure-side permissions / quota availability for the VM family used by the cluster.

In Databricks, I can edit and start the cluster, so the workspace permissions seem fine. The issue appears during provisioning on Azure: the cluster fails when trying to acquire instances for Standard_DC4ads_v6 in North Europe. So my question is whether this is due to Azure VM family quota / capacity / subscription configuration, rather than anything inside Databricks itself.

If I am understanding correctly, I would really appreciate confirmation on whether I need an Azure quota increase or Azure-side enablement for that VM family. Thanks again for your time and availability.