How to handle Sematext becoming detached from monitoring a restarted container - node.js

Question

I am using Sematext to monitor a small composition of Docker containers plus the Logsene feature to gather the web traffic logs from one container running Node Express web application.
It all works fine until I update and restart the web server container to pull in a new code build. At this point, Sematext Logsene seems to get detached from the container, and so I lose the HTTP log trail in the monitoring. I still see the Docker events, so it seems only the logs part which is broken.
I am running Sematext "manually" (i.e. it's not in my Docker Compose) like this:
sudo docker run -d --name sematext-agent --restart=always -e SPM_TOKEN=$SPM_TOKEN \
-e LOGSENE_TOKEN=$LOGSENE_TOKEN -v /:/rootfs:ro -v /var/run/docker.sock:/var/run/docker.sock \
sematext/sematext-agent-docker
And I update my application simply like this:
docker-compose pull web && docker-compose up -d
where web is the web application service name (amongst database, memcached etc)
which recreates the web container and restarts it.
At this point Sematext stops forwarding HTTP logs.
To fix it I can restart Sematext agent like this:
docker restart sematext-agent
And the HTTP logs start arriving in their dashboard again.
So, I know I could just append the agent restart command to my release script, but I am wondering if there's a way to prevent it from becoming detached in the first place? I guess it's something to do with it monitoring the run files.
I have searched their documentation and FAQs, but not found anything specific about this effect.

Answer 1

I seem to have fixed it, but not in the way I'd expected.
While looking through the documentation I found the sematext-agent-docker package with the Logsene integration built-in has been deprecated and replaced by two separate packages.
"This image is deprecated.
Please use sematext/agent for monitoring and sematext/logagent for log collection."
https://hub.docker.com/r/sematext/sematext-agent-docker/
You now have to use both Logagent https://sematext.com/docs/logagent/installation-docker/ and a new Sematext Agent https://sematext.com/docs/agents/sematext-agent/containers/installation/
With these both installed, I did a quick test by pulling a new container image, and it seems that the logs still arrive in their web dashboard.
So perhaps the problem was specific to the previous package, and this new agent can "follow" the container rebuilds better somehow.
So my new commands are (just copied from the documentation, but I'm using env-vars for the keys)
docker run -d --name st-logagent --restart=always \
-v /var/run/docker.sock:/var/run/docker.sock \
-e LOGS_TOKEN=$SEMATEXT_LOGS_TOKEN \
-e REGION=US \
sematext/logagent:latest
docker run -d --restart always --privileged -P --name st-agent \
-v /:/hostfs:ro \
-v /sys/:/hostfs/sys:ro \
-v /var/run/:/var/run/ \
-v /sys/kernel/debug:/sys/kernel/debug \
-v /etc/passwd:/etc/passwd:ro \
-v /etc/group:/etc/group:ro \
-e INFRA_TOKEN=$SEMATEXT_INFRA_TOKEN \
-e CONTAINER_TOKEN=$SEMATEXT_CONTAINER_TOKEN \
-e REGION=US \
sematext/agent:latest
Where
CONTAINER_TOKEN == old SPM_TOKEN
LOGS_TOKEN == old LOGSENE_TOKEN
INFRA_TOKEN = new to me
I will see if this works in the long run (not just the quick test).