Most of the time it seems like connectivity issues like this are due to network configurations, either for the applications themselves or for Docker. Unfortunately there are lots of points of failure between a Docker container in one server and a Docker container on another; someone else might have experienced the same symptoms but with a different cause. Have you tried using any network analysis tools to try and trace where the connection is going awry?
For example, you could test the connectivity between servers by running a copy of netcat (aka nc) on each machine. The machine that would normally run InfluxDB would run netcat on port 8086 using the
-l argument, which listens for connections on an incoming port. Then you could test the connection from the machines which run Telegraf and Chronograf by connecting with their own copy of netcat.
You could follow a similar process at each step of the network hop, and narrow down the problem. With more specific details and diagnostic data, people on these forums will be better able to help you figure out what’s going on, or you might find a fix on your own.
Other tools that might be of interest are ngrep and tcpdump. If you’re on Linux with a recent kernel, eBPF might be of interest. Brendan Gregg has a blog post about eBPF tracing tools.