PMM Server /ping Bad Gateway

We have a PMM server instance (v1.17.2) running as a docker container on an EC2 instance. Unfortunately, we have some issues regarding linux metrics. It seems to be the case that the PMM server is misconfigured, but I’m unfortunately unable to find the exact issue.

The web dashboard is working fine, registering new client with pmm-admin add also works without problems. But when running pmm-admin check-network on the clients, we get the following message:

[B]* Connection: Client <-- Server[/B]

-------------- -------- ----------------- ------- ---------- ---------


-------------- -------- ----------------- ------- ---------- ---------

linux:metrics worker1 <ip>:42000 [B]DOWN[/B] [B]YES[/B] [B]YES[/B]

AWS SG are setup correctly. Interesstincly, when calling

curl http://pmm-server-ip/ping

we get a

502 Bad Gateway

Also, when running

pmm-admin list

on the client it shows the service linux metrics as running, but also displays

Error getting a list of external metrics: status code 502 (text/html).

Please check versions of your PMM Server and PMM Client.

I’ve check the containers nginx

cat /var/log/nginx/error.log

which shows a lot of error messages like this one:

2019/11/26 10:54:55 [error] 155#0: *1 connect() failed (111: Connection refused) while connecting to upstream, client:, server: _, request: "GET /qan-api/ HTTP/1.1", upstream: "", host: ""

What is going on here? Is the server misconfigured? Thanks in advance for anyone helping out!