Explore strategies for reliable data processing with minimal manual intervention in this 30-minute conference talk from SREcon21. Gain insights from Google and Slack experts on mitigating risks in data processing pipelines. Discover techniques for implementing safer batch jobs, including canarying, automated global rollouts, and various validation methods. Learn how these approaches can significantly reduce the manual work required when updating batch jobs across millions of users. Understand the importance of these practices in maintaining efficient and reliable data processing systems at scale.
Overview
Syllabus
SREcon21 - Reliable Data Processing with Minimal Toil
Taught by
USENIX