Textfile collector log spam

[mttradebyte]

Host operating system: output of uname -a

Linux xxxxxxx 4.9.93-41.60.amzn1.x86_64 #1 SMP Fri Apr 13 21:58:27 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

node_exporter version: output of node_exporter --version

node_exporter, version 0.16.0 (branch: HEAD, revision: d42bd70)
build user: kitchen@b8f2acb37099
build date: 20180613-11:13:36
go version: go1.10.3

node_exporter command line flags

--collector.textfile.directory=/opt/node_exporter/textfiles --collector.filesystem.ignored-mount-points ^/var/lib/kubelet/.*

Are you running node_exporter in Docker?

No

What did you do that produced an error?

Nothing, merely running as usual

What did you expect to see?

No errors in the logs

What did you see instead?

Huge number of errors in the logs

We've noticed that we're getting a load of spam in our logs regarding metrics retrieved from a text file in the text file collector directory. We were originally running node exporter 0.15.2, however after the resolution of #704, I compiled and upgraded to 0.16.0 in the hope it would resolve our issue but it appears not. We're still getting messages akin to the below every time node_exporter runs, which at 15s is resulting in millions of errors per week clogging up /var/log/messages.

node_exporter: time="2018-06-13T12:53:44Z" level=error msg="error gathering metrics: 6 error(s) occurred:\n* collected metric process_cpu_seconds_total counter:<value:109.93 > was collected before with the same name and label values\n* collected metric process_max_fds gauge:<value:4096 > was collected before with the same name and label values\n* collected metric process_open_fds gauge:<value:496 > was collected before with the same name and label values\n* collected metric process_resident_memory_bytes gauge:<value:8.03487744e+08 > was collected before with the same name and label values\n* collected metric process_start_time_seconds gauge:<value:1.528870600633e+09 > was collected before with the same name and label values\n* collected metric process_virtual_memory_bytes gauge:<value:3.382833152e+09 > was collected before with the same name and label values\n" source="log.go:172"

[SuperQ]

This is not really a bug or problem with the node_exporter. It indicates you have invalid data in your textfiles.

[mttradebyte]

This is one of the particular metrics it's complaining about:
# HELP process_virtual_memory_bytes Virtual memory size in bytes.
# TYPE process_virtual_memory_bytes gauge
process_virtual_memory_bytes 3.382833152E9

All of these metrics were exported using the jmx_exporter javaagent, whitelisting only java.lang:*

[SuperQ]

If you have the jmx_exporter agent, Prometheus should scrape those directly. The node_exporter is not designed to be a proxy for other exporters.

[SuperQ]

For usage questions/help, please use our community. There are many more people there that can help. GitHub issues are reserved for bug reports and feature requests.