Large XML Reports cause crash due to OOM #192
Description
Background
When fetching a report that is large (e.g. 500MB+) CxFlow can throw out of memory exceptions and crash. This is more common in memory constrained environments such as Github Actions where 7GB of memory is available for the task, even when that available memory is allocated for JVM heap space via JVM flags.
Suggested Solution
The reports need to be processed in a more memory efficient manner to prevent these crashes so that large reports will not crash CxFlow.
Details
Based on several stack traces, the issue occurs within getReportContent here: https://github.com/checkmarx-ltd/checkmarx-spring-boot-java-sdk/blob/develop/src/main/java/com/checkmarx/sdk/service/CxService.java#L381
The XML report is treated as a String type and several operations are performed on the string creating several pieces of garbage for the GC. When the report is large, these pieces of garbage are equally large and will quickly consume all available heap space.
The XML report is read into memory as a string:
https://github.com/checkmarx-ltd/checkmarx-spring-boot-java-sdk/blob/develop/src/main/java/com/checkmarx/sdk/service/CxService.java#L395
This line creates two pieces of garbage by trimming the string twice. This commonly triggers the OOM error.
In other log examples, this function also causes OOM:
https://github.com/checkmarx-ltd/checkmarx-spring-boot-java-sdk/blob/develop/src/main/java/com/checkmarx/sdk/service/CxService.java#L402