Showing results for 
Show  only  | Search instead for 
Did you mean: 

AWS Fargate best deployment option


Hi All,

Please can some one help me to understnad best approach to monitor AWS Fargates out of three approaches mentioned in DT docs , fargates running on ECS service.

Currently we are using runtime injection, just want to understand Bulidtime injection has any advantages over runtime.


Dynatrace Mentor
Dynatrace Mentor

Hi @gurugogi ,

Thank you for your question. The best approach really depends on your circumstances and what works for you, since each environment is different. 

Build time injection would simply be done once and less at deployment time, while the runtime injection involves automated download of agent files. In addition, build time injection is a less overhead intensive way to build the image, but the trade-off is the increase in size for application image and build time. The main two differences between the two are installation and set up related.

Dynatrace ACE Services Consultant


Thanks @Taylor-Sanchez .

what I understood is for "runtime injection" will have two containers, one is application container and another one is Oneagent container, oneAgent container will have required code modules based on the technology and inject it on to application container, injection will happen only time when these containers are run for first time, so oneagent container will run for onetime and inject code modules onto application container and then it will be in stopped status.

Please let me know what I understood is correct or not.

This is the path on oneAgent container which is available on volume mount where OneAgent code modules are stored ?
Scroll to Environment, and in Environment variable, define LD_PRELOAD with the value /opt/dynatrace/oneagent/agent/lib64/

DynaMight Legend
DynaMight Legend

This is a very interesting topic. I've seen issues with monitoring Fargate. The main issue is that the Fargate node will not have access to download the Oneagent and do auto injection. Which then causes a gap in the monitoring stack. This is just one of the issues I've run into with Fargate. All the other K8 connections I've seen and work with were pretty seamless, Fargate is the one that gave us the most problems. 


Featured Posts