Using PHP to process 1000s of API events per second per machine and millions of events per minute is not simply a matter of tuning PHP. This talk will cover proven, battle-tested strategies for building, testing, monitoring and scaling PHP APIs. This talk will cover system tuning, PHP tuning and response tuning that apply to any API, as well as Lumen tuning, and will also cover specific capabilities in Google Cloud such as auto-scaling, tracing and load balancing. This is not a theoretical talk or a talk not backed by real-world experience. I will be covering specific lessons learned and implemented to process billions of monthly API events. I will cover specific examples, caveats, and alternatives along the way. I've spent the last year scaling the API architecture and code at PubWise to handle millions of events per minute and to auto scale nearly without limit. We've moved through three phases of architecture based on scaling needs and learned a ton in the process. I want to pass on what we've learned.