Platform engineering streamlines infrastructure management, enhances developer productivity, and ensures reliability. By treating internal software development and DevOps teams as customers, platform engineers can enhance developer productivity, increase DevOps automation, and optimize cloud costs.
However, simply establishing a platform team and hoping to reap the benefits isn’t guaranteed success. In the worst case, you end up with gatekeepers for certain tools that get overwhelmed with support tickets. To ensure the platform team can add real value and make the lives of all developers in your organization easier, they must follow best practices. Automation and self-service portals are the keys to success.
This article explains those practices after a small refresher on platform engineering.
What is platform engineering?
Platform engineering is a software engineering methodology that builds tools and services for internal developers. Its goal is to centralize and automate recurring DevOps and platform ops tasks, such as infrastructure provisioning and APIs or generating SDKs, to prevent duplication of efforts between multiple teams.
Platform engineering teams enhance developer experience by creating an internal developer platform that provides access to all resources they need, without requiring ticket requests or other overhead that takes away from development pipelines.
For example, the an internal developer platform makes reusable and standardized software components accessible to developers through self-service portals, while providing DevOps teams the tools to manage those components continuously. The platform centralizes the management of complex infrastructure while decentralizing access for engineering teams organization-wide.
As a result, the development team and operations team can both benefit from improved productivity, cloud security, and cost efficiency.
For a more in-depth discussion on platform engineering and its importance, check out this article on our blog.
Platform engineering best practices
Now that the definition is clear, let’s discuss platform engineering best practices for DevOps teams.
Sticking to these recommendations will guide your platform engineering practice so you don’t get lost in needless discussions or pour work into features nobody will use.
Establishing a platform engineering strategy
Building as you go might be a good approach for software projects where potential outcomes are initially diffuse, but platform engineering benefits from a clear path. You know your main goal is improving developer productivity, but it’s important to define tangible steps so your team knows what to do next at any point in time.
Identify the key metrics that define your success and the roles and responsibilities not only for your platform teams, but also the DevOps and IT teams who will act as stakeholders.
Understanding platform engineering tools and resources
Do your research to ensure you know what tools and resources your developers and operations team are using, and why they’re using them. Otherwise, you might design and implement your internal developer platform based on incorrect assumptions, and your engineers won’t use it. Also, some developers are attached to their tools, and integrating them might be better than outright replacing them.
To obtain this data, evaluate your current deployments and interview engineering teams about the tools they prefer to use.
Estimating developer productivity gains
Understand the manual tasks across your software development and operations team. Identify what slows them down, what’s error-prone, and what could most benefit from automation.
Engineers hate recurring manual tasks, and machines are usually better at doing these repeated tasks without any deviations. If you understand what causes delays or increases frustration for your teams, automating those tasks will give you the biggest gains and loyal promoters across the organization.
If you know what your engineers are doing, you can prioritize automation for manual tasks that take up most of their time.
Defining cloud governance standards
Standardization is paramount for platform engineering, and there’s nowhere that’s more important than in the cloud. The more work developers can do with one type of resource, the easier the platform will be to manage. Ungoverned cloud activity creates security risks and drives up unnecessary costs.
Determine the resources your teams need, then create policy files to define activity that will be prohibited. For cloud governance, an internal developer platform provides a valuable point of control for infrastructure provisioning. This allows you to democratize infrastructure provisioning as long as it adheres to your governance standards, while also providing the assurance that nothing your teams deploy violates them.
Some examples include:
- Cloud platforms, regions, or services: Create a policy to allow only specific infrastructure technologies to ensure your teams align with your standards
- Instance sizes, environment runtimes, and maximum costs: Custom policies allow you to prevent wasted cloud costs before they happen.
- Private cloud storage: Rely on the internal developer platform to identify any public cloud storage buckets before they go live to prevent a common but potentially devastating security risk proactively.
For DevOps teams, this approach democratizes access to infrastructure provisioning and saves time that go toward more important tasks.
Evaluating security and governance implications
Role-based access controls (RBAC) and encryption for secrets management enable a platform engineering team to democratize cloud infrastructure provisioning while reducing access to sensitive information.
Define roles for access permissions, account credentials, and security keys for your teams’ cloud environments. The internal developer platform can manage all secrets, remove them from the provisioning process, and enforce RBAC to improve security along with developer productivity.
Last but not least, log developer activity to ensure you know where your permissions are lacking and to give you insights into how engineers are using your platform.
Automating routine Day-2 actions
Automate, automate, automate! Similarly to DevOps, it’s the main mantra of platform engineering.
The less time spent on a task, the more tasks your teams can handle, and the more likely you are to eliminate human errors due to distraction or boredom.
Automating day-2 actions such as patches, health checks, and vulnerability scans to keep the maintenance burden low.
Implement self-service solutions to enhance developer experience
Support tickets can bring your platform team to a screeching halt, so you should prioritize self-service solutions wherever possible. You don’t win anything by giving your platform engineering team more work— the workload must be reduced.
Provide your users with service catalogs so they can find the resources they need without submitting a ticket. Give them contextual suggestions when using the platform so they don’t get lost. Finally, let your teams share knowledge via collaboration tools: community support can be a strong enabler for spreading know-how across your organization.
Offer documentation and training materials for your tools and services to help developers understand how to use the platform effectively. Regular workshops or training sessions can also enhance adoption and proficiency.
Implementing feedback loops
Continuously collect developer feedback to ensure you’re on the right track. You can do this with feedback forms or recurring interviews, but it can also be done automatically through activity logs and performance monitoring. Analyze this data to understand how well your platform meets developer needs and where there’s room for improvement.
Reviewing resources for optimization
You’re not finished with the first release—from there you must iterate repeatedly to improve the platform and allow it to react to market changes.
Review and optimize the resources allocated to different teams regularly. This can involve right-sizing cloud resources to prevent over-provisioning, create cost efficiencies, and meet performance demands.
Platform engineering is a powerful method for elevating your DevOps strategy to the next level. Treating your internal engineers like customers, and providing them with state-of-the-art tooling and services, empowers them to perform at their best.
Your platform team must standardize and automate as much work as possible and stay in tune with its users while leveraging self-service techniques so they don’t get overrun by support tickets.