London DevOps #86 - AWS

** Please bring photo ID for entry **

London DevOps returns to AWS at Principal Place with three talks lined up - please see the provisional agenda below:

6:00pm - Arrival

6:45pm - Introductions

7:00pm - The Talks

Talk 1 - Incident Groundhog Day - Hamed Silatani, CEO of Uptime Labs

Learning how to respond effectively to incidents is hard. One of the reasons is that we never see the same incident twice. While we can learn vital lessons during and after an incident, we can’t hop into a time machine, and apply these lessons to the same incident to discover their impact. What if we could experience the same incident over and over again? What might we learn? This talk describes a ‘staged world’ experiment in which 20 incident managers separately experienced the same simulated incident affecting a fictitious e-commerce company. We discuss what we noticed that differentiated some incident responders from others, and some surprising things that we expected to see, but didn’t.

Talk 2 - PlatformOps and Wardley Maps - Andrey Adamovich, CTO | Software Delivery Coach | IT Strategy

Internal platforms are crucial for empowering software teams, but their evolution presents strategic challenges. In this session, we’ll embark on a journey using Wardley Maps to explore the platform engineering landscape together. Rather than offering definitive answers, we’ll engage in a discussion to map out key elements, track their evolution, and uncover potential strategies. Through this collaborative process, we aim to spark new ideas and perspectives, enabling participants to better navigate the complexities of building and scaling internal platforms.

Talk 3 - Scaling Stateful Workloads In The Cloud - Roman Gershman, Co-Founder & CTO at DragonflyDB

Designing applications as horizontally scalable, stateless workers is often seen as the easiest path to achieving near-infinite scalability. However, this approach hinges on the ability to offload state management to external systems, typically databases or in-memory datastores. But what happens when your datastore itself becomes the bottleneck? Why do some of your cluster shards hit 100% CPU utilization while others remain underutilized?
In this talk, we'll explore the challenges of scaling the datastore layer, particularly in cloud environments, where specific factors can significantly impact your cluster’s performance. We'll delve into the trade-offs between vertical and horizontal scaling, and discuss strategies to overcome common bottlenecks.

9:00pm - Pub

Once the meetup is finished, we will head to Beechwood Sports Bar & Kitchen for a few drinks and networking.

Participate in a future Meetup. If you'd like to speak at a future meetup, or if you are able to host or sponsor the event, please fill in this form and we'll get back to you.