- Frame: Rails 5.0.0.1
- Platform: Heroku
- Server: Puma, 30 processes with 10 workers each
We observe an increase in memory once per hour, coinciding with the Ruby garbage collector, as seen in the screenshot below. The number of requests per unit of time almost constantly increased in memory (~ 1300 rpm).
The memory seems stable, with the exception of garbage collector assemblies, usually fluctuating several megabytes in any direction around a fairly stable average. Debugging the application locally using profiling tools, such as a memory profiler or dumping heap space using object allocation tracing, did not ultimately identify memory leaks.
Question:
How do I know if this is related to the garbage collector?

source
share