The previous post in the series can be found here.
Time flies. Over 4 months have passed since the last post about our automation journey - there are so many other topics to cover that the posts related to testing saga kept getting postponed ... It may seem like there's probably nothing to write about, but such a statement would not be true. However, keeping in mind the continuous, progressive nature of the automation, it's not that easy to distinguish key moments & game-changing milestones.
Anyway, I've decided to conclude the series today. Why so? We're getting THERE. We've set our course, survived quite a few storms & ... it works for us, we already see the light in the end of tunnel.
At the time of writing these words:
- our automation artifacts cover twenty-few percent of all carefully designed tests cases (E2E) for the whole platform
- new approach (JavaScript + Cypress, involving Front-End Engineers to support QA) works like charm - no-one has decided to leave the QA team because of Project Kalashnikov: people have embraced the challenge & the general belief in a successful automation (& reaching Continuous Delivery) is now higher than ever
- our QA crew has encountered and tackled (with a full support of other engineers) several additional challenges on the way: test stability ("blinking"), multi-repo VS single-repo, test running parallelism, test result reporting, splitting suite into parts based on criticality, etc.
- teams pro-actively organize internal case-writing sessions where experienced Front-End engineers share their practices to help everyone else write better, more efficient & more effective tests
- standard pre-release regression test period was shortened twice already (in general: by about 40%)
What's more important - the trend is very optimistic. The goal is not getting away, it's getting closer. More and more within reach. And people see it, very clearly. Of course there'll be more bumps - but these days no-one treats them as excuses or showstoppers; rather obstacles to be crushed/walked around w/o even slowing down.
It all has happened not due to heroic effort of any single individual, but thanks to collaboration of a wider group of people who've started to believe they can do it together. To get things automated (concluded fact), instead of just automating (activity in progress, without the indication of an end) things.
It took few moments of hardship to bind them & there's still work to be done here, but ... damn ... I'm proud of them & of the stuff they've done already. Keep it up guys, this simply can't fail now.
