Closed katopz closed 7 years ago
The address should be localhost:9090 since that's the port Prometheus is listening to. 9091 is the mapping for the requests coming from outside the container. Since the target looks for localhost, it should be the internal port 9090.
I did notice the same error before but it disappeared a bit later when it tried to scrape it again. Can you double check that's not your case?
I rerun that stack a few minutes ago and go everything green. If in your case prometheus endpoint continues being red after a while (give it a few minutes just in case), we should try to figure out what's the difference between my system and yours.
For some reason it's appear all green now after remove all node and redo all over again, I can't tell what go wrong there because I try too many thing at that time even reinstall docker and also restart.
One thing I notice is Docker somehow provide different IP after many remove all nodes and retry, usually it always like...
node-1 = 192.168.99.100
node-2 = 192.168.99.101
node-3 = 192.168.99.102
but somehow it was like...
node-1 = 192.168.99.106
node-2 = 192.168.99.107
node-3 = 192.168.99.108
And start to have this issue (and solved by that comment) in between.
I'll close this in case for now in case I can't reproduce it any more.
Thanks
After deploy prometheus-grafana-df-proxy.yml
I found that at page
http://localhost:9091/targets
has prometheus endpoint
http://localhost:9090/metrics
but it link to
http://b1b1472b3e3e:9090/metrics
which look like node id instead of node ip.And actually it should be
http://localhost:9091/metrics
? (It did print output fine at 9091)Thanks