Datadog-Container-ID header is not sent

[ddyurchenko]

Hi guys,

I hope everything is going great for you! :)

I want to report a minor issue that I caught. I was testing nginx + cpp application inside Docker and found that Datadog-Container-Id header is never sent to /traces for opentracing-cpp instrumentation. According to Containers Tagging RFC https://github.com/DataDog/architecture/blob/master/rfcs/apm/agent/containers-tagging/rfc.md, it should be sent every time traces are sent.

Please feel free to contact me if you need any additional details.

[dgoffredo]

This doesn't look too difficult.

It might be covered by telemetry work that we've discussed doing (but haven't scheduled) to collect information like this, kubernetes service, executable name, etc.

I'll mention it at the next standup meeting.

[paddycarey]

Curious if there's been any progress on this one? We're in the middle of an EC2->ECS migration for our NGINX instances and the lack of infrastructure-related tags on our NGINX traces is a problem.

I'm not an experienced C++ dev but happy to take a stab at this if it's desirable.

[dgoffredo]

We haven't looked at this, sorry. The telemetry work that I mentioned has stalled, too.

It probably won't be a while until I or @cgilmour have the bandwidth to look into container ID detection in the tracer. Instead, is there a feature of the Datadog Agent that can discover the nginx container's ID instead? I looked over the RFC linked above, and it doesn't look like it.

Also, what's the plan for dealing with cgroup v2?