cancel
Showing results for 
Search instead for 
Did you mean: 
Warehousing & Analytics
Engage in discussions on data warehousing, analytics, and BI solutions within the Databricks Community. Share insights, tips, and best practices for leveraging data for informed decision-making.
cancel
Showing results for 
Search instead for 
Did you mean: 

Can not create a Streamlit Databricks App on free tier

carlaoftw
New Contributor
I'm trying to deploy a minimal Streamlit app as a Databricks App on a fresh Free/Community workspace and I get:

"Compute error - App creation failed unexpectedly. Please remediate by deleting the app."

As alternatives, I also tried the API and the CLI and had the same results.
Am I missing something in the setup, or is this a bug or service limitation on the Free/Community tier?
I also attached a screenshot of the UI error.

Thanks for your time and support.
1 ACCEPTED SOLUTION

Accepted Solutions

Louis_Frolio
Databricks Employee
Databricks Employee

Hey @carlaoftw ,  thanks for sharing the screenshot—this helps confirm where the failure is happening.

 

What’s going on

  • Databricks Apps are supported on Free Edition (the new free tier that replaces Community Edition) with quotas, including “one app per account.” If you’re on Free Edition, you aren’t hitting an unsupported-tier issue by trying to deploy an app.
  • The banner “Compute error — App creation failed unexpectedly. Please remediate by deleting the app.” indicates the app’s serverless compute couldn’t be provisioned, not a problem with your Streamlit code or app.yaml/requirements.txt per se.
  • There has been a recent service-side issue seen by the Apps on-call team where app creation fails during the internal “create_compute_resources” step and surfaces exactly this error message; one root cause was a regional backend failure (gatekeeper not returning a cluster tier for certain Free/Pharos workspaces). In those windows, delete-and-retry won’t help until the region recovers.

Is this a limitation of the free tier?

  • No—Apps are allowed on Free Edition. Limits include “one app per account” and automatic stop after 24 hours, but the feature itself is available. If you had already created an app in the account, you would see a quota/limit error rather than this compute creation error.
  • Do note: Apps aren’t supported on Standard-tier paid workspaces. If you try a paid trial, make sure it’s Premium/Enterprise, not Standard; otherwise Apps won’t work by design.

Quick checks to rule out local setup issues

  • Verify you have no other app in the account (Free Edition limit = 1). If one exists (even stopped), delete it before trying again.
  • Ensure today’s Free Edition credits/quotas haven’t been exhausted; Free Edition is serverless-only and enforces daily usage caps which can prevent new compute from starting if you’ve hit the limit.
  • Keep your app small: large files (>10 MB) in the source code directory cause deployment failures; minimal Streamlit apps should be well under this.
  • Minimal template reminder (matches what the UI walkthrough says): create app.yaml that specifies the startup command plus environment variables, and requirements.txt for pip dependencies; then deploy via UI/CLI. Example shown in the product’s inline guide on the page you captured.
    bash # Example CLI flow databricks apps create --name dashboard-app --source-code-path . databricks apps deploy dashboard-app --source-code-path .

Practical ways to proceed

  • Retry later; if this is the regional compute-provisioning issue, it usually resolves when backend services recover.
  • If you need to move forward immediately, try:
    • Creating a new Free Edition account (which may land in a different region), then deploy the same minimal app.
    • Using a paid workspace on a supported tier (Premium/Enterprise) and region; avoid Standard-tier workspaces as Apps aren’t supported there.
  • When it does deploy, remember the Free Edition “one app” quota and the 24-hour auto-stop behavior so you can plan restarts and avoid hitting limits unexpectedly.
 
Hope this helps, Louis.

View solution in original post

5 REPLIES 5

Louis_Frolio
Databricks Employee
Databricks Employee

Hey @carlaoftw ,  thanks for sharing the screenshot—this helps confirm where the failure is happening.

 

What’s going on

  • Databricks Apps are supported on Free Edition (the new free tier that replaces Community Edition) with quotas, including “one app per account.” If you’re on Free Edition, you aren’t hitting an unsupported-tier issue by trying to deploy an app.
  • The banner “Compute error — App creation failed unexpectedly. Please remediate by deleting the app.” indicates the app’s serverless compute couldn’t be provisioned, not a problem with your Streamlit code or app.yaml/requirements.txt per se.
  • There has been a recent service-side issue seen by the Apps on-call team where app creation fails during the internal “create_compute_resources” step and surfaces exactly this error message; one root cause was a regional backend failure (gatekeeper not returning a cluster tier for certain Free/Pharos workspaces). In those windows, delete-and-retry won’t help until the region recovers.

Is this a limitation of the free tier?

  • No—Apps are allowed on Free Edition. Limits include “one app per account” and automatic stop after 24 hours, but the feature itself is available. If you had already created an app in the account, you would see a quota/limit error rather than this compute creation error.
  • Do note: Apps aren’t supported on Standard-tier paid workspaces. If you try a paid trial, make sure it’s Premium/Enterprise, not Standard; otherwise Apps won’t work by design.

Quick checks to rule out local setup issues

  • Verify you have no other app in the account (Free Edition limit = 1). If one exists (even stopped), delete it before trying again.
  • Ensure today’s Free Edition credits/quotas haven’t been exhausted; Free Edition is serverless-only and enforces daily usage caps which can prevent new compute from starting if you’ve hit the limit.
  • Keep your app small: large files (>10 MB) in the source code directory cause deployment failures; minimal Streamlit apps should be well under this.
  • Minimal template reminder (matches what the UI walkthrough says): create app.yaml that specifies the startup command plus environment variables, and requirements.txt for pip dependencies; then deploy via UI/CLI. Example shown in the product’s inline guide on the page you captured.
    bash # Example CLI flow databricks apps create --name dashboard-app --source-code-path . databricks apps deploy dashboard-app --source-code-path .

Practical ways to proceed

  • Retry later; if this is the regional compute-provisioning issue, it usually resolves when backend services recover.
  • If you need to move forward immediately, try:
    • Creating a new Free Edition account (which may land in a different region), then deploy the same minimal app.
    • Using a paid workspace on a supported tier (Premium/Enterprise) and region; avoid Standard-tier workspaces as Apps aren’t supported there.
  • When it does deploy, remember the Free Edition “one app” quota and the 24-hour auto-stop behavior so you can plan restarts and avoid hitting limits unexpectedly.
 
Hope this helps, Louis.

carlaoftw
New Contributor
Hello Louis, I hope this message finds you well!

To add a bit more context to this post, I was aware of the free account limits regarding the number of apps and how long they can run before being shut down.
To deploy these apps, I was using the “Custom (from scratch)” option from the UI to avoid any code-related issues, but I was still getting the same error.
The error also occurred when I tried a simple Streamlit “Hello World”.

Today, the applications started working as expected.
So, as you said, it was probably a compute provisioning issue happening in my region.

I had already created a second account (in the same region) to check if the issue was related to a faulty account creation, but that wasn’t the case. I was a bit skeptical about a service-side compute provisioning problem because it affected my accounts for about three days, and I hadn’t seen any reports similar to mine.

Thank you for all the information.

Cordially,
Carlos

eebruno
New Contributor II

I'm facing the same problem.

DataBeli
New Contributor

I am facing the same problem now

eebruno
New Contributor II

Good morning, everyone.

The app is working again. I've already tested it and it's working.