How do I run a node container on AWS ECS without exiting - node.js

Question

I'm struggling to keep my node.js container running on ECS. It runs fine when I run it locally with docker compose, but on ECS it runs for a 2-3 mins and handles a few connections (2-3 health checks from the load balancer), then closes down. And I can't work out why.
My Dockerfile -
FROM node:6.10
RUN npm install -g nodemon \
&& npm install forever-monitor \
winston \
express-winston
RUN mkdir -p /usr/src/app
WORKDIR /usr/src/app
COPY package.json /usr/src/app/
RUN npm install
COPY . /usr/src/app
EXPOSE 3000
CMD [ "npm", "start" ]
Then in my package.json -
{
...
"main": "forever.js",
"dependencies": {
"mongodb": "~2.0",
"abbajs": ">=0.1.4",
"express": ">=4.15.2"
}
...
}
In my docker-compose.yml I run with nodemon -
node:
...
command: nodemon
In my cloudWatch logs I can see everything start -
14:20:24 npm info lifecycle my_app#1.0.0~start: my_app#1.0.0
Then I see the health check requests (all with http 200's), then a bit later it all wraps up -
14:23:00 npm info lifecycle mapov_reporting#1.0.0~poststart: mapov_reporting#1.0.0
14:23:00 npm info ok
I've tried wrapping my start.js script in forever-monitor, but that doesn't seem to be making any difference.
UPDATE
My ECS task definition -
{
"requiresAttributes": [
{
"value": null,
"name": "com.amazonaws.ecs.capability.ecr-auth",
"targetId": null,
"targetType": null
},
{
"value": null,
"name": "com.amazonaws.ecs.capability.logging-driver.awslogs",
"targetId": null,
"targetType": null
},
{
"value": null,
"name": "com.amazonaws.ecs.capability.docker-remote-api.1.19",
"targetId": null,
"targetType": null
}
],
"taskDefinitionArn": "arn:aws:ecs:us-east-1:562155596068:task-definition/node:12",
"networkMode": "bridge",
"status": "ACTIVE",
"revision": 12,
"taskRoleArn": null,
"containerDefinitions": [
{
"volumesFrom": [],
"memory": 128,
"extraHosts": null,
"dnsServers": null,
"disableNetworking": null,
"dnsSearchDomains": null,
"portMappings": [
{
"hostPort": 0,
"containerPort": 3000,
"protocol": "tcp"
}
],
"hostname": null,
"essential": true,
"entryPoint": null,
"mountPoints": [],
"name": "node",
"ulimits": null,
"dockerSecurityOptions": null,
"environment": [
{
"name": "awslogs-group",
"value": "node_logs"
},
{
"name": "awslogs-region",
"value": "us-east-1"
},
{
"name": "NODE_ENV",
"value": "production"
}
],
"links": null,
"workingDirectory": null,
"readonlyRootFilesystem": null,
"image": "562155596068.dkr.ecr.us-east-1.amazonaws.com/node:06b5a3700df163c8563865c2f23947c2685edd7b",
"command": null,
"user": null,
"dockerLabels": null,
"logConfiguration": {
"logDriver": "awslogs",
"options": {
"awslogs-group": "node_logs",
"awslogs-region": "us-east-1"
}
},
"cpu": 1,
"privileged": null,
"memoryReservation": null
}
],
"placementConstraints": [],
"volumes": [],
"family": "node"
}
Tasks are all stopped with the status Task failed ELB health checks in (target-group .... Health checks pass 2 or 3 times before they start failing. And there's no record of anything other than an http 200 in the logs.

Answer 1

I was using an old version of the mongo driver ~2.0, and keeping connections to more than one db. When I upgraded the driver, the issue went away.
"dependencies": {
"mongodb": ">=2.2"
}
I can only assume that there was a bug in the driver.