Part 1: Permission management for databases

Permission management for databases is a sore spot in many DevOps pipelines. 

It requires a careful balancing act between access control and productivity. On one hand, privileged access exposes the organization to risks. On the other hand, if we restrict it too much, we end up with bottlenecks resulting in a lack of productivity. It’s a lose-lose situation. 

It’s Important to Limit Direct Access

Companies strive to limit direct access concerning permission management for databases for many reasons, such as security, as a fail-safe against potential human errors and a number of other benefits. No matter the reason, this type of least-privilege access policy has led to a rise in popularity for BI tools such as BigQuery and Elasticsearch. These tools allow devs the opportunity to access the data without entering the production environment.

However, Situations Exist When Access Needs to be Direct

Although it’s prudent to follow the zero-trust approach, there are still times when it’s necessary for engineers to enter the production environment immediately, for example, during incidents. Without immediate access to fix things in production, problems can persist and the MTTR rises, resulting in lost time, lost resources and lost money. 

A few of examples are the following:

• Incidents
• Supporting customers
• Maintenance and implementations
• Integrations
• Eliminating silos between developers and operations

Benefits of Automating Role and Permission Management for Databases

It’s clear there needs to be a solution to the chaos, and there are two approaches to achieve this. They are the following:

1. One option is to transform it into a makeshift project that requires constant manual maintenance, involving the granting and revoking of permissions every time an individual joins or departs from the company, for example. This would result in a complicated and time-consuming endeavor, where a significant amount of time is wasted setting up permissions for all users. Clearly, this falls short of the ideal solution.

2. The other option involves employing a flexible and fully automated system for role-based permission management that enables you to maintain a suitable balance between collaboration and control. This approach provides you with peace of mind, knowing that your data is secure and protected.

Problems with Manual Management of Database Permission Management

1. Lack of Visibility. Apono provides centralized identity visibility so that teams can easily see each identity’s entitlements and permissions across all platforms. With this information, they can make informed decisions about which action to take.

2. Not Secure. Standing permissions can introduce significant risk if not revoked when they are no longer needed, since they would provide an attacker with access if a user’s credentials are stolen. For organizations seeking to minimize their attack surface, mitigate the risk of data breaches, and ensure compliance, it is crucial to prioritize the reduction of accounts with standing privileges and transition towards a zero standing privilege framework.

3. Bottlenecks. While we may have succeeded in decreasing complexity and increasing security by removing all access, we inadvertently introduced a bottleneck, restricting the number of individuals who can perform tasks on the database. The absence of self-service capabilities for developers will impede the release process, as it will rely on manual intervention from the database administrator (DBA) to implement necessary changes. This problem gives rise to numerous productivity challenges and effectively halts the CI/CD (Continuous Integration/Continuous Deployment) process whenever database-related tasks are involved.

4. Static. Different situations call for different solutions, and permission management is no different. With dynamic authorization, you specify the authorization conditions and behaviors that govern each of your resources, including data stores, APIs, and applications. Conventional role-based access control (RBAC) lacks precision in the authorization process, while dynamic authorization offers a more intricate authorization service. Instead of relying on fixed permissions and role assignments to safeguard your resources, you can now configure policies that consider a wide range of attributes. By incorporating external attribute data, you can make highly detailed authorization decisions, placing a greater emphasis on context in the decision-making process. This context is evaluated for every resource request, enabling a tailored approach to authorization management that surpasses the coarse-grained nature of traditional authorization.

Benefits of Automated JIT Access to Databases

Just-in-Time access grants users just in time and just enough access to mission critical and sensitive databases. It eliminates the need for long-term privileged access for users and reduces the attack surface.

1. Centralized Visibility. Database permissions are usually messy. We have zero visibility and too many passwords and permissions, making it difficult to know who has permissions to what. Having a centralized depository that automatically documents and manages user permissions is essential for post-mortem analysis.

2. Time-restricted access. Automating the revoking of permissions after a specified time period assures there are no standing permissions that could be abused. 

3. Tool Adoption. Company-wide tool adoption is just as important as what the tool does. If it’s not user-friendly, there will be friction. It’s important the tool fits in smoothly with the rest of the tool stack.

It is evident that when it comes to permission management for databases, we must establish order amidst chaos, and luckily there are ways to create a win-win permission management situation when it comes to provisioning database access.