Galaxy APM memory leak

I never experienced this type of behavior until I switched over to Galaxy’s APM (was previously using Kadira), but my memory usage is continuously increasing over time. I have had to log in periodically to kill the current container and reset the memory consumption - anyone else experience this since switching to Galaxy APM?

I’ve been noticing this behavior in my app even without Galaxy’s APM or Kadira. Does Meteor / Node leak memory?

We need to have this solved as my app has been restarting automatically a few times per day due to “out of memory” error, which is not good for user experience or data integrity.

It is of course possible that there could be a memory leak in meteor or node; however it’s much more plausible that the leak is in package or application level code. Check your server code, then see which packages you are using and see if they might be the culprit. I’ve just removed react-router-ssr from a project running on galaxy as I suspect a memory leak.

1 Like

Looks like there is an issue with Node 4.8.1, which was fixed in Node 4.8.2, which will be updated in Meteor in the next patch release: https://github.com/meteor/meteor/issues/8630

3 Likes

Thanks for pointing that out. I’ll wait patiently and then update my app in production as soon as Meteor 1.4.4.2 is released.

1 Like

See my post in this thread: Please help me troubleshoot: Application exited with signal: killed (out of memory)

1 Like