Anaconda Notebooks FAQ#

General FAQ#

What are notebooks and why would I use them?#

Jupyter Notebooks provide a web-based interface for creating and sharing computational documents. You can seamlessly mix executable code, documentation, and instructions in one portable document. Notebooks are not only a great portable learning tool, but also a highly capable vehicle for prototyping and producing data science work.

Anaconda Notebooks lets you skip setup and installation and get straight to learning and writing code.

How do I access Anaconda Notebooks?#

You can access and use Anaconda Notebooks from any modern web browser and anywhere you have an internet connection.

After you have logged into your account on Anaconda Cloud, go directly to nb.anaconda.cloud or click on “Notebooks” from the top navigation bar of Anaconda Cloud.

What do I have access to?#

With Anaconda Notebooks, you get all of the following running on our resilient and supported cloud platform, so you can use it anywhere on any device!

Features

Free

Starter

Pro/Business

A dedicated JupyterLab notebook interface

✓

✓

✓

Fast, backed-up SSD storage

5GB

10GB

20GB

CPU seconds (daily)

1,000

4,000

8,000

Published applications

1

2

4

Conda environments with the most popular python packages

✓

✓

✓

Ability to create and upload your own custom environments

✓

✓

✓

Example notebooks

✓

✓

✓

Is Anaconda Notebooks different from Jupyter notebooks?#

Anaconda Notebooks is a hosted JupyterLab service that enables you to run JupyterLab notebooks reliably online. Your dedicated JupyterLab instance comes preconfigured with persistent cloud storage, hundreds of data science packages, and a managed infrastructure.

What are the pros and cons of publishing on Anaconda Notebooks versus working on PyScript.com directly?#

Publishing on Anaconda Notebooks provides you with a server-hosted app, while PyScript.com provides you with a browser-hosted app. Panel supports both server and browser operation, but a) browser-side operations require copying all the data down to the browser (not suitable for very large datasets), and b) not everything can be run browser-side because not every operation is available in WASM (e.g. libraries like numba, dask, or pytorch cannot be run in the browser currently). In other words, it’s a matter of running on the server or running locally in your browser.

Where can I get support?#

You can get community support on the Anaconda Community forums. If you’re in need of further technical assistance, please file a support ticket.

What packages are preconfigured on Anaconda Notebooks?#

All packages available from the Anaconda installer are preloaded and ready to code through Anaconda Notebooks. More specifically, the service will include environments based on the most recent installers. For example, anaconda-panel-2023.05-py310 is the latest release of Anaconda Distribution and is the default environment within Anaconda Notebooks. As new installers are released, new environments will be available.

To see a list view of all preloaded packages, launch Anaconda Notebooks and select the anaconda-panel-2023.05-py310 kernel. Once the kernel is activated, enter conda list into any notebook file.

Can I share my notebooks?#

Yes! Click Share at the top of your notebook to produce a shareable link or embeddable HTML for your notebook. See Sharing Anaconda Notebooks for more information.

How do I upload a notebook to the service?#

In the Anaconda Notebooks JupyterLab interface, click Upload files in the File Browser to browse for a local .ipynb file. Then, click Open. The notebook will appear in the left-hand menu.

You can also drag and drop a notebook from a folder on your system to the file browser to upload it.

How do I save a notebook?#

Like most IDEs or editors, JupyterLab has the standard “Save” and “Save As…” functions that will save a notebook in your directory on our platform. You can also download a notebook file from the File menu to save it locally.

What kind of storage does Anaconda Notebooks come with?#

The storage provided through the notebook service is persistent Elastic Block Store (EBS) storage. EBS storage is fast, backed-up, SSD storage and supports common data science and machine learning workloads. EBS storage is generally faster and more reliable than most cloud-hosted options.

Can I add more storage?#

Not yet, but soon! If you’re running out of storage space, we suggest that you remove any unused notebook assets, such as extra file directories, notebook files, and custom conda environments.

What are the memory limits of this service?#

On this service, each process is limited to 6GB of memory. If you exceed that, your process will be killed and you will need to restart your kernel. If you need to run much larger processes, please contact us at sales@anaconda.com.

What is a high-compute second?#

A CPU second is one second of running code on a single CPU core at 100%. We refer to them as “high-compute seconds” on our pricing page to clearly distinguish CPU seconds from “wall clock” seconds. Simply running JupyterLab, writing code, and using the interface don’t really use up quota (though they have a small impact). Only running python code from within a notebook and running commands from the terminal count against your quota, and even then very few command functions truly tax the CPU.

For example, if your code makes an HTTP request, then it will use a tiny amount of CPU time assembling the request and sending it out over the network, but will then use no CPU at all while it’s waiting for a response. When the response comes back from the other end, then it will again use a small amount of CPU to interpret the response and provide your code with the results. So, in general, CPU time is only used while your program is actively making calculations, not while it is waiting for other systems.

When does the clock on CPU seconds reset?

Our notebook service accounts have a per-day limit for the maximum number of seconds fully utilizing the CPU. Once an instance hits that limit, it is not shut down, but instead given lower CPU priority and a limit to the amount of compute resources available. This limit is reset every day, so full compute access will be restored the next day.

Can I use packages from the Professional repository in Anaconda Notebooks?#

Packages available from Anaconda Notebooks are a subset of packages available from the free and public repo.anaconda.com repository. Installing packages from the Professional repository via tokenized access is not currently supported.

Can I install new packages or create custom environments in Anaconda Notebooks?#

You can create your own conda environments using any packages that conda can install from repo.anaconda.com. This can be achieved by following the steps in Anaconda Navigator’s Managing environments documentation, or via the command line interface (CLI):

Tip

These steps can also be found in the README.ipynb file in your Anaconda Notebook.

Creating custom environments

  1. Open a terminal from the Launcher in Anaconda Notebooks.

  2. Run the following command to create a custom environment:

    # Run this command to create a custom environment running Python 3.9
# Replace <ENVIRONMENT_NAME> with a name of your choosing
conda create --name <ENVIRONMENT_NAME> python=3.9 ipykernel -y

Activating custom environments

After a minute or two, you should be able to activate your custom environment by either:

  • Clicking the kernel at the top right of the notebook (“anaconda-<YEAR>.<MONTH>-py<PYTHON_VERSION>”), then switching to the kernel of the environment you created in the Select Kernel modal.

  • Selecting the notebook displaying your custom environment name from the Launcher.

Installing packages

You can then install any further packages you need by running the following:

# Replace <PACKAGE_NAME> with the name of the package you want to install
conda install <PACKAGE_NAME> -y

Note

Custom environments will be stored using your dedicated, persistent Anaconda Notebooks storage. This ensures the custom environment will be available after the current session.

Can I use Anaconda Notebooks for work?#

Customers accessing Anaconda Notebooks with subscription tiers Pro and above are permitted to use all Anaconda products for commercial use. However, Anaconda Notebooks alone does not provide commercial compliance to its users.

I have an organization in Anaconda Cloud. How can my team leverage Anaconda Notebooks?#

Registered customers who are part of organizations on Anaconda Cloud can independently access Anaconda Notebooks. Access to Anaconda Notebooks is granted upon member role designation and registration.

Can I control access to Anaconda Notebooks?#

All registered customers can access Anaconda Notebooks. Organization-level features, including user access controls, are coming soon. Stay tuned!

I have a site license. How do I give my members access?#

If you are a customer but have not yet registered your organization on Anaconda Cloud, please refer to this documentation on how to set up your organization and invite members.

How do I create an R kernel?#

Open a terminal from the Launcher in Anaconda Notebooks and run conda create -n test_r r-irkernel -y. The kernel should appear within a few minutes.

Troubleshooting#

How do I completely reset my notebook instance?#

To completely reset (“factory reset”) your instance of Anaconda Notebooks, email user care at usercare@anaconda.com.

My notebook is trying to import a package, but I’m getting an error.#

The most common cause of errors is a lack of required package(s) installed in your environment. The default environment we provide, based on the Anaconda distribution, contains hundreds of the most common python packages for data science, but it doesn’t include everything. You may need to create a custom environment to install the package you need.

Here are a couple of steps to help resolve this:

Make sure you have the right kernel/environment selected

The default anaconda-<YEAR>.<MONTH>-py<PYTHON_VERSION> environments have a broad selection of packages, but you may have created a custom environment for your notebook. Separate environments are represented as “kernels” in JupyterLab. You can view and switch between available kernels by clicking the kernel name in the upper-right corner of the content pane.

List the packages available in an environment

You can view which packages are available in your current environment from the terminal by running the conda list command. If you want to view the packages of a specific environment, run the command conda list -n <ENV_NAME>. If you need to see a list of available environments, you can run the conda env list command. An asterisk will appear next to your current active environment.

Tip

You can run those commands directly in a code cell within your notebook just by adding a “!” to the front of the command (e.g. !conda env list).

Create a custom environment

If none of your existing environments have the right package(s), either install the package into one of your custom environments with conda install <PACKAGE> or create a new custom conda environment with the right packages. You can add new environments via the terminal by running conda create --name <ENV_NAME>.

Once an environment is created, it will be available as a kernel for running your notebook.

What can I do if my notebook is running really slowly?#

You may have exceeded your CPU usage limit for the day. Our notebook instances have a limit for the maximum number of seconds fully utilizing the CPU. Once an instance hits that limit, it is not shut down, but instead given lower CPU priority and a limit to the amount of compute resources available. This limit is reset every day, so full compute access will be restored the next day.

To see current progress towards your daily quota, reference the widget in the upper right of the interface that shows current CPU usage vs. the daily limit.

To better manage your CPU usage, regularly check the Running Terminals and Kernels widget in the left sidebar to kill unnecessary kernels when you no longer need them.

What do I do if I run out of storage/go over my quota?#

Caution

Creating custom environments consumes a large amount of storage. Anaconda recommends free tier Notebooks users avoid custom environments.

You can check the status of your disk usage via the widget in the top right of the screen, which shows current usage as a percentage of the total space available.

If you’re running out of space, upgrade your subscription or delete some items from your drive:

Do you have any extra notebooks or directories you can remove?

You can view and delete files from the File Browser in the upper left, or on the command line by launching a terminal.

Do you have any custom conda environments?

  1. Run conda env list and see if there are any environments NOT in /opt/conda.

  2. If there are, you can remove those that you don’t need anymore by running:

    # Replace <ENV_NAME> with the environment name
conda env remove -n <ENV_NAME>

  3. Further, clear out the cache and other artifacts by running:

    conda clean --all
pip cache purge
rm -rf /tmp/*

Note

The option to upgrade your account and expand your storage is coming soon!

Why am I receiving an error message?#

If you receive a “file load error,” “unhandled error,” or “unexpected error,” like in the following figure, you have most likely exceeded the storage space for your current tier. Follow the steps in the storage question above to remove items from your Notebooks instance, or upgrade to a higher subscription tier.

I’m registered/signed in—why isn’t Notebooks opening?#

Your browser’s pop-up blocker (automatically enabled on Firefox and Safari) may have prevented Notebooks from opening.

Disable your pop-up blocker and try opening Notebooks again from anaconda.cloud.

I have upgraded from the free tier to a paid tier, but I am unable to connect to certain websites.#

Unrestricted internet access is only activated in new Notebooks processes. Therefore, Anaconda recommends restarting the kernel or starting a new notebook.

Why does Anaconda use an allowlist?

Anaconda uses an allowlist to prevent malicious actors from using free accounts to hack into and spam other websites anonymously.

Free tier accounts can only access the websites on our allowlist.

Paid tier accounts have unrestricted internet access, as they can be linked to real people via the payment details.

How can I add sites to the allowlist?

To add new sites to the allowlist, submit a request using the Anaconda Notebooks/PythonAnywhere Allow List Request form. We only add sites to the list if they have an official, public, documented API—that is, sites that are designed and intended for machine consumption rather than human consumption.

Note

GitLab instances: GitLab instances can be allowlisted if they contain public repositories. To add a GitLab instance to the allowlist, provide a link to the public repository in your request.

I published a Panel application, but the application is blank.#

There are a couple of reasons your application may not be rendering:

  1. To create a valid Panel application, one or more of your outputs must be marked as .servable(). Take this minimal “Hello, World!” program, for example:

    import panel as pn
pn.Row("Hello, World!").servable()

    If you added content to your application but there’s still nothing showing up, ensure that your notebook can be run from top to bottom. The easiest way to test this is to click Kernel in the menu bar, then select Restart Kernel and Run All Cells.. from the dropdown.

  2. Ensure you have selected the anaconda-panel-2023.05-py310 kernel from the kernel selector in the top-right of your notebook.

I published an application but it’s stuck in a “publishing” state.#

If your application is stuck in the “Your app is being published” state, check your notebook error logs. Address any issues raised and republish.