Data API security
This section covers security related to the Data API.
Last updated
This section covers security related to the Data API.
Last updated
This article covers the security aspects of using the Bubble Data API specifically. If you want to learn more about the Bubble Data API in general, you can check out the articles below:
Article: The Data API Article series: The Bubble API
The Data API allows external apps and systems to interact programmatically with the data stored in your Bubble app. It provides a set of endpoints for creating, reading, updating, and deleting data.
Naturally, giving an external system access to your app's database can open up for potential vulnerabilities, and it's important to learn how to use it correctly.
As we've explored multiple times in our Security article series, also applies strongly to the Data API. We'll now go over how you can plan and execute to adhere to this principle when you set up the Data API.
In a bank, not every employee has the keys to the vault. A teller can access the cash drawer but doesn't have the authority to authorize large wire transfers. Conversely, the bank manager might have that authority but doesn't necessarily need access to every single safety deposit box.
In our Introduction to APIs article we explored the two sibling concepts of authentication and authorization.
The way this logic is structured is no different than in real life: before you let someone in through your front door, you would want to know who it is. In that way, they authenticate (you recognize their face), and this allows you to authorize what they have access to (step into your house).
The same pattern of identification and authorization is applied in countless real-life scenarios:
Airports: (who is the passenger? What plane do they have access to)
Cinemas: (does the person have a ticket? To what movie?)
Banks (who is the customer? What bank accounts do they have access to)
The reason for highlighting this in real-life scenarios is to emphasize the authentication/authorization way of thinking to your security planning. When you are planning security for your Data API, you can ask the question in exactly the same way:
Who is the client trying to access my app's database?
What should this client have access to?
When you combine this way of thinking with the principle of least knowledge, we have some very helpful rules of thumbs to guide your thinking:
Who is the client trying to access my app's database?
I need to know in order to apply the principle of least knowledge
What should this client have access to?
According to the principle of least knowledge – only as much as they need and not an inch more
Having established those principles, let's look at what that means in practice as you develop your app.
The first step in any secure transaction is that of knowing who the client is. For the Data API, this follows a range that we can split into three categories:
No one has access (the Data API is disabled)
Some clients have access (the Data API is enabled, but requires authentication)
Everyone has access (the Data API is enabled, and requires no authentication)
Before you make any decisions regarding your app's Data API, you need to determine whether anyone should have access at all. The Data API is disabled by default, but can be enabled in the Settings – API section of your app by checking Enable Data API.
If your app doesn't need any external app or system to access the database, you can safely keep that checkbox unchecked. This keeps the Data API completely disconnected and no other security measures are needed.
The second level in our access range is to give one or more clients access, but not everyone. This is where most apps with an enabled Data API will be, and it means we need to authenticate ("check the ID") of every system that wants access to the database.
There are two ways to authenticate an external app or system without giving everyone access. To learn how to set up each one, following the article link listed below:
Authenticating as a user essentially means to "log in" to your app in the same way as a regular user would. This means that you can apply Privacy Rules to control what that user has access to. This is an access level that gives you a high degree of control over what that user can and can't do in your database. Article: The Data API - authenticating as a user Using this authentication method, you will usually continue to set up rules for who can access what. Or in other words, authorization. Follow this link to continue setting that up.
Authenticating as an admin means to use Bubble's built-in API token system to give access. This automatically gives them full administrator access, which means the same access level as you have as the app's developer (not to be confused with any admin role you set up in your app). This access level grants extensive permissions, allowing the reading, modification, and deletion of all data in your app. Ensure it's used only with external apps or systems you trust implicitly. Article: The Data API - authenticating as an admin
In this option, we disregard the authentication step altogether: the door is open and anyone can walk in. The logic then follows that since we're not checking who walks through the door, we can't determine what each visitor should have access to. Everyone is treated the same.
This method should only be used when you are determined to give free, public access to your database. Keep in mind this can not only affect your security, but also the amount of that your app consumes (as API requests will spend server resources, adding to your total workload).
To learn how to set up your Data API to offer everyone access, see the article below: Article: The Data API - no authentication
Returning to the bank metaphor: a person who registers as a customer in a bank needs to have access to their own account, but most would find it strange if they were given access to all accounts. The same logic applies to an app where you don't want to share all data even to people that are authenticated to use the app in the first place.
We've taken care of the authentication (who is the client), and now it's time to explore how we determine the authorization for those client (what the client can access). This is done in two ways:
The first step to protect data is to enable only the data types that you want to expose.
Keep in mind the following:
Unchecked data types are not available in the Data API regardless of how the user authenticates
Checked data types are exposed, but adhere to the privacy rules in combination with the client’s authentication. We'll look at data types in the next step.
When you enable the Data API, you will see the list of of data types that you want to show.
The data types that are not enabled (unchecked) will not be accessible by anyone. In the example above, we have a data type Product that we want to share with another app. Following the principle of least privilege, we'll only activate that one, and keep all other data types unchecked.
This setting takes care of the overarching authorization: what data types any client has access to. In the next step we'll look at privacy rules: they provide the fine-grained control
After enabling the data types we want to expose, it's usually a good idea to use privacy rules to control which fields on those data types can be accessed, and what kind of actions can be performed.
Just like in your regular app, privacy rules are set up on a user-basis. In other words, which user you authenticate determines the level of access they have to:
Find in searches
What fields and uploaded files to expose
Make changes to
Create
Edit
Delete
For more in-depth instructions on how to set up the privacy rules for a data type exposed in the API, check out the article below:
Article: The Data API - Setting up Privacy Rules
With the methods above in mind, let's go over the tools that secure your Data API.
Keep the principle of least access in mind (link)
