Current Setup:

1. Java application running as Kubernetes workload on Google Kubernetes engine
2. JVM args : "-XX:+UseContainerSupport", "-XX:NativeMemoryTracking=detail", "-XX:+HeapDumpOnOutOfMemoryError", "-XX:HeapDumpPath=/dumps/oom.bin", "-Xmx300m", "-Xms300m"

3. Docker memory set using deployment.yml

   limits: cpu: 350m memory: 700Mi requests: cpu: 200m memory: 128Mi

4. Native memory committed is below 550m when the pod is very near to hitting the max container limit. This leads me to believe that there is something apart from native memory causing this issue

 Total: reserved=1788920KB, committed=546092KB
    -                 Java Heap (reserved=307200KB, committed=307200KB)
                                (mmap: reserved=307200KB, committed=307200KB) 
     
    -                     Class (reserved=1116111KB, committed=75431KB)
                                (classes #12048)
                                (malloc=1999KB #25934) 
                                (mmap: reserved=1114112KB, committed=73432KB) 
     
    -                    Thread (reserved=46444KB, committed=46444KB)
                                (thread #46)
                                (stack: reserved=46240KB, committed=46240KB)
                                (malloc=153KB #270) 
                                (arena=51KB #88)
     
    -                      Code (reserved=259362KB, committed=57214KB)
                                (malloc=9762KB #14451) 
                                (mmap: reserved=249600KB, committed=47452KB) 
     
    -                        GC (reserved=1034KB, committed=1034KB)
                                (malloc=26KB #180) 
                                (mmap: reserved=1008KB, committed=1008KB) 
     
    -                  Compiler (reserved=321KB, committed=321KB)
                                (malloc=189KB #982) 
                                (arena=133KB #5)
     
    -                  Internal (reserved=39461KB, committed=39461KB)
                                (malloc=39429KB #16005) 
                                (mmap: reserved=32KB, committed=32KB) 
     
    -                    Symbol (reserved=15588KB, committed=15588KB)
                                (malloc=13214KB #128579) 
                                (arena=2374KB #1)
     
    -    Native Memory Tracking (reserved=3220KB, committed=3220KB)
                                (malloc=249KB #3553) 
                                (tracking overhead=2971KB)
     
    -               Arena Chunk (reserved=178KB, committed=178KB)
                                (malloc=178KB)

5. Have also checked memory consumption of java process using ps (550m) and htop (res 604m)

Issue

The pod has single container and that container has only this java application running with above mentioned configs. This java application essentially downloads images in parallel using threads with some validations like image size, dimension etc. The pod memory consumption keeps increasing until it almost hits that limit and the application restarts. I have read through many different posts and articles and I can't figure out what is causing this memory to keep going up.

Edit 1: Docker image is based off openjdk:8-jdk-slim Edit2: Some screenshots of monitoring