Performance Testing at Postmark

Performance testing at Postmark involves keeping a history of all performance tests, tracking the health of the system, and using a combination of load, spike, and stress testing. The team uses their staging environment for most tests, with isolated tests performed in production only when necessary. Their workflow includes loading test on staging, performing spike or stress tests, comparing results to stable branch, repeating until better or same as stable branch, loading test on production, and tracking system health over 24 hours. Key points of interest include email content size, type, number of attachments, sending API/SMTP limits, maximum number of senders, sending frequency, total emails sent, ramp up period, time to send emails, response times for each email, average response times, validation related metrics such as email response messages and statistics from every status update. The team aims to measure everything possible, find the lower limit at which performance degrades, test on staging environment, thoroughly perform tests before releasing to production, monitor performance after release, track history of performance tests, and start with simple tests and expand accordingly.