TechPrufer
Get started
Atlassian
Atlassian logoAtlassian · Feature trail

Atlassian Cloud Engineering Overview

1 post · Apr 2021 – Oct 2024

What this trail covers

Atlassian's cloud platform, Micros, has evolved its multi-account strategy to address AWS API and resource limits. Initially operating within three large AWS accounts, the platform experienced significant rate-limiting issues impacting deployment speed and reliability. The 'Account Sharding' project incrementally addresses this by introducing new accounts across development, staging, and production environments. This strategy aims to gradually reduce the number of services per account, enhance scalability, and enable future account-per-service models. The migration focuses on compute workloads, with careful consideration for ephemeral and persistent resources, and prioritizes minimizing disruption to service owners.

All Atlassian trails
  1. From firefighting to future-proofing: Atlassian’s journey with a multi-account strategyThis post details the 'Account Sharding' project, a re-design of Atlassian's internal cloud platform (Micros) to address scaling challenges caused by AWS account limits. It explains the impact of hitting soft and hard AWS resource and API limits, illustrating the problem with CloudFormation API call graphs for development, staging, and production environments. The post outlines the decision-making process for account strategy, opting for an incremental approach of adding new accounts (4x capacity increase initially) rather than a full migration or account-per-service model. It also discusses the migration effort for different resource types (compute, ephemeral, persistent) and the decision to avoid direct service owner involvement in migrations.
  2. Atlassian’s Cloud Engineering OverviewThis post serves as an introductory overview of Atlassian's cloud engineering landscape. It outlines the core components and philosophies that underpin their product development and operations, including their technology stacks, architectural patterns (like ARIs), orchestration system (Micros), communication mechanisms (StreamHub), observability tools (Splunk, SignalFx, LightStep, Sentry, Pollinator), build/deploy infrastructure (Spinnaker, Artifactory, LaunchDarkly), and data services (Socrates). It also briefly mentions front-end and mobile platforms, cross-product experiences, and the Forge development platform.