Atlassian DevOps / SRE Interview Questions

1.Tell me about a time you quickly prototyped an infrastructure solution to solve an immediate production problem. What did you build and how did you validate it worked?

2.Tell me about a time you were using infrastructure tooling daily and discovered a significant pain point that led you to build or advocate for improvements. What did you find and what happened?

3.Walk me through the last time you had to coordinate a major infrastructure change across teams in different time zones. How did you ensure everyone stayed aligned without constant meetings?

4.Describe a monitoring or alerting practice you introduced to your team that significantly improved how you handle incidents. What was broken before and how did you get team buy-in?

5.Walk me through the riskiest infrastructure experiment you've run in production in the last two years. What made it risky, and how did you manage to ship it safely while learning quickly?

6.Describe the most complex cross-team infrastructure migration you've led with teams spread across multiple continents. How did you maintain technical coordination and team morale when people rarely overlapped in working hours?

7.Your team deploys Jira microservices to Kubernetes clusters across multiple AWS regions. During a recent deploy, one region's pods started OOMKilling while others were fine. How would you debug this and prevent it from happening again?

8.You're tasked with reducing the blast radius of deployments for Bitbucket's Git hosting service without slowing down the release velocity that development teams expect. What architectural and operational changes would you implement?

9.You're designing a CI/CD pipeline for a new Confluence feature that needs to deploy to 30+ microservices. The product team wants fast feedback on feature flags, but you need to ensure changes don't break existing functionality. What's your approach?

10.Atlassian's Elasticsearch clusters are approaching capacity limits due to increased usage of Atlassian Intelligence features. You need to scale them without impacting search performance across Jira and Confluence. Walk me through your scaling strategy.

11.Design a monitoring and alerting system for Loom's video processing pipeline that handles millions of async video uploads daily. The system needs to detect issues before they impact user experience while avoiding alert fatigue.

12.Code a Terraform module that provisions a secure, multi-environment Kafka cluster for Atlassian's event streaming needs. Include proper networking, security groups, and cross-region replication setup.

13.Design a backup and disaster recovery system for Confluence that can restore any page version within 5 minutes, handles 100TB of content, and needs to work during complete AWS region outages. What's your approach?

14.Design a deployment orchestration system for Atlassian's microservices that can safely roll out changes across 200+ services while allowing teams to maintain their own release schedules. How do you handle dependency conflicts and rollback scenarios?

15.Bitbucket's Git hosting serves millions of repositories with teams pushing code 24/7. Design a system to automatically scale Git storage and compute resources based on usage patterns while keeping costs predictable.

16.You're building a real-time presence system for Confluence that shows who's currently viewing or editing each page. The system needs to handle 50 million concurrent users across thousands of pages. How do you design this?

17.Design a feature flag service that every Jira request hits to determine UI variations and feature availability. It needs sub-10ms latency and 99.99% uptime even when its database is completely down.

18.Atlassian Intelligence needs to process and index content from every Jira issue, Confluence page, and Trello card to power AI search and summarization. Design a system that can handle 10 billion documents with real-time updates while keeping AI processing costs manageable.

19.Tell me about a time you had to convince skeptical developers to adopt a new infrastructure practice or tool. How did you approach the rollout and what resistance did you encounter?

20.Describe a time when you took initiative to improve a broken process or practice in your team, even though it wasn't your explicit responsibility. What did you change and how did you get others on board?

21.Describe a production incident where you had to coordinate response across multiple teams in different time zones. How did you ensure effective handoffs and maintain incident transparency?

22.Walk me through a time when you identified a systemic infrastructure issue affecting product quality, but the fix would require significant engineering investment. How did you build the case for prioritizing it?

23.Tell me about a time you had to influence a senior engineer or tech lead to change their approach to infrastructure or tooling when you didn't have formal authority over them. What was your strategy?

24.Tell me about the most complex technical decision you've had to communicate to non-technical stakeholders where you needed their buy-in for infrastructure changes. How did you make the case effectively?

25.A customer reports that Jira Cloud is taking 8 seconds to load their project dashboard, but only for projects with more than 1000 issues. How would you systematically investigate and prioritize what might be causing this performance degradation?

26.Estimate how many API calls Confluence receives per day globally. Walk me through your reasoning and assumptions, then tell me how you'd validate if your estimate is reasonable.

27.Our Kubernetes clusters are showing 40% higher memory usage after the latest Bitbucket release, but CPU utilization is normal. No code changes touched memory management directly. How would you approach diagnosing this issue?

28.Atlassian Intelligence is launching a new AI summarization feature that will generate summaries for every Jira issue and Confluence page. Estimate the computational cost of this feature for our first month, assuming we have 500 million documents. What are the key cost drivers you'd monitor?

29.You're investigating why Trello board load times increased by 200ms over the past month, but there were no major feature releases or infrastructure changes. The increase affects all users globally. How do you systematically identify what changed?

30.Estimate the storage cost impact if we enabled unlimited file attachments in Jira issues. Our customers currently upload 10GB per month on average with our 2GB per issue limit. Walk through your assumptions about how usage patterns would change and what it would cost Atlassian annually.