Legacy systems are everywhere. They are difficult to deal with and everybody wants to avoid or replace them. Replacing or rewriting is usually not realistic in the short term and the end result is fire-fighting and slow productivity when dealing with them.
This session targets developers or leaders who are responsible for systems in production. Main focus is Java/JVM systems but other technologies can apply the strategies discussed here. Following topics will be discussed:

• How to decide where to start, getting most of the results with the minimum effort.
• What does fast mean for different people?
• APMs and profilers, which metrics to look for?
• Branching strategies for incremental improvements.
• Minimizing the fear of deployment: how to recover quickly from mistakes.
• Reducing latency of database transactions under high throughput.
• Multithreading: when to do it, and how to do it easily.
• Flamegraphs as a secret weapon for bottleneck discovery.
• Teaching your team and colleagues.
• Cases