Upgrade from .net8.0 to .net10.0 causing periodic 139 AWS ECS Fargate error with log "terminate called after throwing an instance of 'Js::JavascriptException'"

[coultonluke]

Describe the bug

I have been running .net 8.0 for years on AWS ECS without issue. I have updated to net 10.0 just updated the packages that were required as per the upgrade instructions excluding entity framework.

dotnet outdated -mv 10.0 --exclude pomelo --exclude entity --include Microsoft.AspNetCore --include Microsoft.Extensions --include System.Net.Http.Json --include system. -u

I'm now periodically getting containers that are exiting on AWS with "Exit code: 139." where it logs a single line and then the container becomes unresponsive and shortly afterwards deprovisioned. Apparently 139 is a SIGFAULT or error accessing memory?

"terminate called after throwing an instance of 'Js::JavascriptException.

I'm running it in AWS ECS as root and using mcr.microsoft.com/dotnet/sdk:10.0 as the base image.

I'm not sure that I'm doing anything exotic. I'm using ReactJS.net which is an older library and it uses Chakracore-linux which is what handles Javascript, but why would it have an issue every now and again just cause it's a new version of .NET?

I'm building it within codepipeline using docker build.

Unfortunately this is stopping me from upgrading to .net10.0. I have a feeling it's something to do with Chakracore to generate SSR react. Grok says switch to Jint, not sure if that's the answer here or whether it's due to the switch from Debian to Ubuntu.

Thanks in advance for any insight from anyone.

Which .NET image(s) are you using?

mcr.microsoft.com/dotnet/sdk:10.0

Steps to reproduce

Build using docker file, push to AWS ECR and run deploy to ECS. This is my docker file:

FROM mcr.microsoft.com/dotnet/sdk:10.0 AS build-env
EXPOSE 80
WORKDIR /app

# Copy csproj and restore as distinct layers
COPY *.sln .
COPY MySite.WebUI.Core/*.csproj ./MySite.WebUI.Core/

RUN dotnet restore

# Copy everything else
COPY . ./

FROM 369074612400.dkr.ecr.us-east-1.amazonaws.com/node:20.11.1-buster as node-env
WORKDIR /app

# Run webpack
COPY --from=build-env /app ./

WORKDIR /app/MySite.WebUI.Core
RUN npm ci
RUN npm run release
RUN ls -R -1 ./wwwroot/dist

FROM mcr.microsoft.com/dotnet/sdk:10.0 AS second-build-env
WORKDIR /app
COPY --from=node-env /app .
RUN dotnet publish MySite.WebUI.Core -c Release -o out

# Copy static files to S3
FROM 369074612400.dkr.ecr.us-east-1.amazonaws.com/amazon/aws-cli:2.15.28 AS copy-static-content-env
ARG AWS_DEFAULT_REGION
ARG AWS_CONTAINER_CREDENTIALS_RELATIVE_URI
ARG AWS_SECRET_ACCESS_KEY
ARG AWS_ACCESS_KEY_ID
ARG AWS_SESSION_TOKEN
WORKDIR /app
COPY --from=second-build-env /app/out ./out
RUN ls -1 -R ./out/wwwroot/dist
RUN aws s3 cp ./out/wwwroot/dist s3://MySitelive/dist --recursive

# Build runtime image
FROM mcr.microsoft.com/dotnet/aspnet:10.0 AS website
WORKDIR /app
ENV ASPNETCORE_HTTP_PORTS=80
COPY --from=copy-static-content-env ./app/out .
USER root
CMD ["dotnet", "MySite.WebUI.Core.dll"]

Other information

No other issues other than containers just disappearing due to this issue.
Error message is '"terminate called after throwing an instance of 'Js::JavascriptException'"' logged in cloudwatch before the container is deprovisioned.
This happens once every couple of hours and causes hundreds of failed requests due to the container breaking, the application must be shutting down after this error.
Running in AWS ECS Fargate platform version 1.4.0 - Linux
Monitoring shows that I am NOT reaching a max memory limit, the max memory gets to about 26% across all of my containers

Output of docker version

Output of docker info