We’ve had issues in the past with Python autopush with the glib’s malloc implementation. It’s often recommended to use alternative malloc implementations for long running server side programs, such as jemalloc or possibly Google’s tcmalloc.
We can attempt to support running syncstorage under an alternative allocator, then run it on prod’s canary to see if it helps our memory usage.
We’ve had issues in the past with Python autopush with the glib’s malloc implementation. It’s often recommended to use alternative malloc implementations for long running server side programs, such as jemalloc or possibly Google’s tcmalloc.
We can attempt to support running syncstorage under an alternative allocator, then run it on prod’s canary to see if it helps our memory usage.
┆Issue is synchronized with this Jira Task