What are the typical causes of goroutine leaks in production systems and how do they affect the runtime behavior of the service?

Goroutine leaks in production systems usually occur due to:

• Blocking on channels or mutexes, when a goroutine is waiting for data or resource release, but it doesn't happen.
• Waiting on unclosed channels, for example, a goroutine reads from a channel that no one closes or writes to.
• Incorrect use of contexts (context.Context) — if a goroutine does not respond to context cancellation, it may hang in waiting.
• Long or infinite loops without exit, especially if there are no exit conditions or timeouts.

Impact on runtime:

• Gradual increase in memory and CPU resource consumption.
• Increase in the number of active goroutines, which can lead to performance degradation.
• Possible deadlocks and service hangs.

For diagnostics, profiling (pprof), tracing, and monitoring the number of goroutines are useful.