November 2021 Summaries

Yann has joined incident.io as the Head of Sales after co-founding Revive, a mentorship startup. Previously, he worked in Enterprise Sales at Salesforce and transitioned from Solution Engineering. With experience in both startups and large corporations, Yann understands the impact incidents have on companies regardless of size. He enjoys running by the canal between Angel and Victoria Park, surfing, and watching movies. Yann is an avid networker and can be contacted through Calendly or LinkedIn for further information.

At incident.io, we handle sensitive customer data and prioritize its safety in our Go web application. We use multi-tenancy with a single database and application to manage customer data effectively. To ensure data security, we have implemented robust automated testing, organization scoping at the API layer, enforcing organization scope on database interactions, and writing safe-by-default interfaces for caching services. These strategies help us maintain confidence in our ability to keep customer data secure while continuing to develop a great product.

The text discusses the process of reducing Docker image build times from around 10 minutes to approximately 3 minutes for a software development company. It explains how the company's CI/CD pipeline relied heavily on Docker layer caching, but inconsistent performance led them to investigate further. They discovered that CircleCI's Docker Layer Caching feature was not effectively sharing cached layers across multiple build steps due to issues with networked filesystems. To resolve this issue, the company provisioned a dedicated machine with an attached SSD and enough capacity for all their build needs. This allowed them to store all build layers on one large, fast disk. They then connected from CircleCI using SSH to remotely access the Docker builder. The results were significantly faster builds, with the worst builds finishing in under 2 minutes. The author concludes by stating that while this solution may not be perfect for everyone, it has more mileage than most will realize and is a worthwhile investment if it helps the development team move faster.

The text introduces a new member joining the incident.io team as their first non-engineer. This individual will be focusing on Strategy and Ops, with responsibilities ranging from setting up internal processes to developing long-term goals for the company. They bring an eclectic background in various roles across public and private sectors, including Product Manager and Head of Programme for a government initiative. In their personal life, they enjoy practicing yoga, cooking, and spending time by the sea.

Incidents provide valuable opportunities for engineers to expand their knowledge and skills by exposing them to unfamiliar areas of expertise. They often occur at the edges of teams, allowing individuals to learn about topics outside their day-to-day roles. By observing how experienced colleagues handle incidents, engineers can gain insights into effective problem-solving strategies and develop a deeper understanding of their system's infrastructure. Additionally, participating in incidents helps engineers appreciate the importance of writing code that fails safely and utilizing appropriate observability tools to quickly diagnose issues. Lastly, engaging with diverse teams during incident response fosters connections with talented professionals who can offer valuable advice and support throughout one's career.

The text discusses the implementation of logs and tracing in both local development and production environments using Google Cloud Trace and StackDriver. It highlights how adding trace URLs to log lines can significantly improve the debugging experience for developers working locally, allowing them to use familiar production tools while building proficiency with these tools. By integrating traces into their workflow, the team has achieved increased productivity and efficiency in both local development and production environments.

At incident.io, feature-flagging with LaunchDarkly has been instrumental in enabling fast and efficient deployment of features while maintaining control over what is presented to customers. By using a third-party feature management platform-as-a-service like LaunchDarkly, the team was able to seamlessly integrate feature flags across their backend and frontend systems within just one day. This approach has allowed them to continue moving quickly in development while ensuring that only fully tested features are made available to customers. As the complexity of their organization and customer base grows, they can easily add custom attributes for users and use more advanced flag types if needed. Overall, feature-flagging with LaunchDarkly has proven to be a valuable tool for incident.io's development process.

The text introduces a new engineer joining the team at incident.io. They bring web expertise from their previous role as a senior web engineer and Web Discipline Lead at Monzo, where they worked on projects like business banking for customers and internal tooling for customer services. Their passion lies in accessibility, ensuring that the evolving web app remains accessible to all users. In their personal life, they enjoy making music with SHE Choir London and baking tasty treats. They also have a website (localghost.dev) where you can learn more about them and find additional blog posts.