Memory Leak Introduced by `6.12.0-sidecar` release (or other change on Feb 20th 2023)

[PeterCat12]

Hello,

We currently use the ECS Fargate side-care deployment as outlined here.

The specific tag of the lacework/datacollector image we use is lacework/datacollector:6.1.2-sidecar. Since Feb 20th we have noticed a memory leak introduced into all of fargate services that use the lacework sidecar that causes degraded app performance before eventually causing our containers to fall over. This occurs even for services that have NOT been restarted and/or redeployed. Removing lacework from our ECS task resolves the memory leak and prevents any degradation of performance.

1. Am I right in understanding that the agent on our services auto-updates and that the publishing of the latest linux image could have introduced a memory leak into services that haven't even been restart/redeployed?
2. We haven't collected enough metrics using the latest release of the linux agent but wanted to check to see if this would resolve our issue.

Thank you

[PeterCat12]

Using the latest image did resolve the memory leak.