hn.today

Give Me Back My Monolith (2019)

craigkerstiens.com9 points0 comments
Screenshot of Give Me Back My Monolith (2019)

Microservices are losing some of their earlier shine, and teams are increasingly questioning whether breaking a monolith into many services is worth the cost. The case for a monolith is not that it always scales better or suits every organization, but that it keeps routine development tasks comparatively simple. A new developer can often get an application, database, and background worker running quickly; a distributed setup may require learning orchestration tools and launching a large cluster before work can begin, creating needless friction, especially for junior engineers.

Splitting an application also makes failures harder to trace: a problem may cross services, queues, and mismatched software versions, where a monolith might offer one stack trace, consolidated logs, and a debugger. Testing grows more involved too, because integration tests need a working version of the whole system; testing services separately risks missing failures between them. These costs can undermine the promised gains in agility, team scaling, performance, and resilience. Many teams, the author observes, are actually constrained by a single-node PostgreSQL database, so their scaling work often belongs at the database level rather than in a wholesale architectural rewrite. Microservices can be worthwhile when their benefits outweigh the operational burden, but a monolith remains a sensible, maintainable choice for many applications.

Read on craigkerstiens.com0 comments on Hacker News

Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.

More in Programming

The daily digest

Today's best Hacker News stories, summarized and screenshotted, one email a day.