How to investigate RSS memory leak?

[mwilniewiec]

Details

I have a problem with one of my microservices that its RSS memory is rising and I don't know what should be a proper way of finding the real problem. I already analyzed heapdumps and nothing can be found as I can see that heap is pretty stable.

process.memoryUsage() is currently showing:

external: 40866815,   (38.97 MB)
heapTotal: 135319552, (129.05 MB)
heapUsed: 114264216,  (108.97 MB)
rss: 1641189376,      (1565.16 MB)

It's a docker running inside of a k8s pod doing quite a lot of things.

• it's a client of RabbitMQ (using amqplib) processing lots of messages
• using prostgres with typeorm
• using redis
• making GRPC calls to other services
• decompressing data using zlib.inflateSync
• lots of other stuff

Heap:

RSS:

In the charts you can see that there are 2 instances of this microservice and both have exactly the same growth in RSS and exactly the same Heap.

Is there any way that I could analyse the other 1.3 GB of RSS memory which is not in the Heap ?

Node.js version

12.14.1

Example code

No response

Operating system

Alpine Linux v3.11 (in a Docker)

Scope

runtime

Module and version

Not applicable.

[preveen-stack]

Take a look at v8-profiler, may be it can help

[preveen-stack]

Also, pls check response to https://github.com/nodejs/help/issues/4148

[mhdawson]

This may also be helpfull - https://github.com/nodejs/node/blob/main/doc/contributing/investigating-native-memory-leaks.md

[github-actions[bot]]

There has been no activity on this issue for 11 months. The help repository works best when sustained engagement moves conversation forward. The issue will be closed in 1 month. If you are still experiencing this issue on the latest supported versions of Node.js, please leave a comment.

[gireeshpunathil]

it is worthwhile to see what is the system characteristics. summary of https://github.com/nodejs/node/issues/31641 in this context is:

rss stays high for ever, if there is no demand for memory in the system

it is a known issue, and the current workaround is to ignore the observation, the growth will subside when there is a demand.

or in other words - the system will adapt to the need, don't measure too much! :)

[github-actions[bot]]

It seems there has been no activity on this issue for a while, and it is being closed in 30 days. If you believe this issue should remain open, please leave a comment. If you need further assistance or have questions, you can also search for similar issues on Stack Overflow. Make sure to look at the README file for the most updated links.

[github-actions[bot]]

It seems there has been no activity on this issue for a while, and it is being closed. If you believe this issue should remain open, please leave a comment. If you need further assistance or have questions, you can also search for similar issues on Stack Overflow. Make sure to look at the README file for the most updated links.