Key insights

• 🔄 Transitioning from monolith to microservices, Benefits of microservices: independent deployment, reducing stepping on each other's changes, Challenges of scaling and deployment with monoliths
• ⚙️ Considerations for adopting microservices: company size and tooling, Complexities of rolling out changes in a global, always-on environment
• 🚀 Developing microservices to help teams move faster initially, Modularization and RPC calls can improve developer productivity and enable independent deployment
• 💼 Challenges faced by DoorDash during the pandemic led to the adoption of microservices, Microservices may introduce technical debt as the system grows
• 🔗 Challenges of managing a distributed monolith, Microservices as a form of technical debt, Socio-technical complexities of microservices, Trade-offs of sharing databases between microservices, Balancing speed and complexity in microservices architecture
• 📈 Move to microservices solved scalability issues, Challenges include complex interactions, error identification, and dependency management, Despite challenges, the move was worth it for the company's growth and learning opportunities
• 📊 The average fan out for a request to Door Dash's front end is about a thousand RPCs due to the domain principle, Programmers often stick to certain principles, such as 100% test coverage, at their detriment, There is a lack of a way to measure the quality of assertions in test coverage metrics
• 🛠️ Advocacy for simpler programming languages, Challenges the industry to produce new abstractions and frameworks, Emphasizes the importance of understanding how software works, Caring about craft and making software better, Dissatisfaction with current practices and the desire to do better

Q&A

• What does the speaker advocate for in terms of programming languages and frameworks?

  The speaker advocates for simpler programming languages and new abstractions and frameworks that offer different trade-offs. They emphasize understanding how software works, caring about craft, and aiming to make software better while expressing dissatisfaction with current practices.

• What socio-technical complexities are associated with microservices?

  Microservices can introduce socio-technical complexities, including the average fan out for a request to DoorDash's front end, which has approximately a thousand RPCs due to the domain principle.

• How did DoorDash navigate the challenges during the pandemic, and what did they adopt?

  DoorDash faced challenges during the pandemic, leading them to rapidly scale and adopt microservices to address the encountered difficulties.

• What are the trade-offs of microservices?

  Microservices involve trade-offs such as technical debt, socio-technical complexities, and the balancing act of sharing databases between services while maintaining speed and managing complexity in the architecture.

• What challenges do microservices introduce?

  Microservices can bring about complex interactions between services, difficulty in identifying the source of errors, and the management of a large number of dependencies despite solving scalability issues.

• How can microservices impact developer productivity?

  Microservices, through modularization and RPC calls, can initially help teams move faster and enable independent deployment. However, as the system grows, they may introduce technical debt.

• What factors should be considered when adopting microservices?

  When considering microservices, it's important to take into account company size and tooling, as well as the complexities of rolling out changes in a global, always-on environment.

• What challenges are associated with scaling and deployment with monoliths?

  Scaling and deployment with monoliths can pose challenges as multiple developers may start stepping on each other, requiring modularization for independent deployment.

• What are the benefits of transitioning from a monolith to microservices?

  Transitioning from a monolith to microservices can lead to independent deployment, reduce stepping on each other's changes, and improve developer productivity through modularization and RPC calls.

• 00:00 The video discusses transitioning from a monolith to microservices, the benefits of microservices, and the challenges of scaling and deployment with monoliths. It also explores the reasons for considering microservices and the impact of company size and tooling. The conversation highlights the complexities of rolling out changes in a global, always-on environment.
• 05:05 Developing microservices can help teams move faster initially, but can lead to technical debt in the long run. When deploying with a single monolithic app, multiple developers may start stepping on each other, leading to the need for modularization. RPC calls from microservices can improve developer productivity and enable independent deployment. DoorDash faced challenges during the pandemic and had to rapidly scale, leading to the adoption of microservices. While microservices may initially boost team productivity, they can introduce technical debt as the system grows.
• 10:24 Discussion on the challenges and trade-offs of microservices, technical debt, and the complexities of managing a distributed system.
• 15:41 The move to microservices solved scalability issues, but also introduced challenges such as complex interactions between services, difficulty in identifying the source of errors, and managing a large number of dependencies. Despite the challenges, it was worth it for the company as it enabled growth and learning opportunities.
• 21:24 The average fan out for a request to Door Dash's front end is about a thousand RPCs due to the domain principle. Programmers often stick to certain principles at their detriment, such as 100% test coverage and the lack of a way to measure assertion quality.
• 26:50 The speaker advocates for simpler programming languages and frameworks that offer different trade-offs than the monolith or microservices approach. They emphasize the importance of understanding how software works, caring about craft, and aiming to make software better. They challenge the industry to produce new abstractions and frameworks to avoid the trade-off between monolith and microservices. They express dissatisfaction with current practices and advocate for doing better.