A developer encountered a peculiar bug where saving new entries sometimes caused duplicate primary key errors. This issue stemmed from the application's custom message ID generator, which used the current timestamp. The generator attempted to handle rapid requests by sleeping for a second if timestamps collided. However, this approach was inefficient and essentially a convoluted autoincrementing counter. A critical flaw was that the previousId was never updated, meaning collisions were not being prevented. The problem was exacerbated because the application was distributed across multiple servers. This distribution meant that even a correctly updated previousId would not prevent ID collisions between different nodes. Potential solutions involve incorporating machine-specific identifiers like MAC addresses or using large random numbers. Alternatively, the developer suggests relying on the database's native auto-increment functionality instead of custom code. The current throughput was low enough that the database could likely manage ID generation effectively.