Though the craze around Pokémon GO has mostly subsided by this point, few will forget the record-smashing launch that it had. Obviously, all those involved with the game's development grossly underestimated how popular it would be, and the game was positively riddled with bugs and instability in the first couple weeks. Now, the servers have largely caught up with the demand, but a Google engineer named Luke Stone recently made a blog post discussing the behind-the-scenes details of what was happening at the time.
At the beginning, the worst case scenario estimates for server strain were put at approximately five times the projected launch target. Within fifteen minutes of the Australian launch, the data already showed that they were going well beyond expectations, and the global numbers eventually rose to fifty times that of the original expectations. The strain was so great that the team made the call to upgrade to a newer version of Kubernetes – a cloud data management platform that Pokémon GO was built on – post-launch, which Stone compared to "switching out a plane's engine mid-flight". You can find the full post here; it makes for quite a fascinating read.
What do you think? Should Niantic have put their expectations higher? How well do you think they handled the overwhelming demand? Drop us a comment in the section below.