PMM2 linux metrics only

it is not a firewall issue

I added an external service (nginx.service)

but maybe I don’t understand pmm2 and make something wrong

Could you please help me with this issue?

Here are examples of configuration commands.

You should use one of them.

Please make sure that it’s shown metrics mode “push” in pmm-admin list output.

I used --metrics-mode=push

but still

Screenshot 2020-12-14 at 16.20.09.png

why?

Could you provide an output of the next command?

Please replace 172.17.0.2 by your PMM server IP.

curl -k https://admin:admin@172.17.0.2/v1/version

Screenshot 2020-12-14 at 18.30.18.png

push mode was not introduced until v 2.12.0 so you will need to upgrade both client and server to take advantage. Since you are using AMI, easiest way to upgrade is using the panel on the main dashboard and allow the packages to update (Amazon marketplace takes a while to approve new AMI’s so they always lag what we have on our download repo).

Another place to check what issues you’re having on the current version is navigate your browser to https://:/prometheus/targets

in <=2.11.1 it’s going to be an HTML output that will show failures from the standpoint of the PMM server, in 2.12.0 it’s text based but you’ll still be able to see errors for any “down” state exporters like:

state=down, endpoint=http://centos.local:42000/metrics?collect%5B%5D=...,
labels={agent_id="/agent_id/59bd4c5e-cfa8-429b-afe0-fc39d6cef1e1",agent_type="mysqld_exporter",instance="/agent_id/59bd4c5e-cfa8-429b-afe0-fc39d6cef1e1",
job="mysqld_exporter_agent_id_59bd4c5e-cfa8-429b-afe0-fc39d6cef1e1_mr-5s",machine_id="/machine_id/e52213f019f94f929ee4895dc9ec97b2",node_id="/node_id/321a3d71-9bb6-4f7e-91bc-cf710f26d322",
node_name="centos",node_type="generic",service_id="/service_id/64639b5d-3481-47bf-9718-e1db62c820ca",service_name="PS5.7",
service_type="mysql"}, last_scrape=4.319s ago, scrape_duration=4.001s, 
error="error when scraping \"http://centos.local:42000/metrics?collect%5B%5D...\" with timeout 4s: timeout"
job="node_exporter_agent_id_9fc926a3-47b3-406c-8c83-3c3c54feb180_hr-1s" (0/1 up)

In that example the client is down but firewalls or unroutable client IPs (from the server perspective) could give the same output. You will likely see some sort of “Context deadline exceeded” (aka: timeout). if you want to post a sample of the error we’ll see what additional help we can provide.

ok, it’s helpful info, but I have another problem.

Percona image from AWS

I can’t update pmm to 2.12 version