Checking VictoriaMetrics health endpoint returns socket timeout

[matejzero]

Hello,

I'm trying to monitor Victoria Metrics health endpoint, but check_http is returning CRITICAL - Socket timeout.

Output of the command:

./check_http -H localhost -u /health -p 8428 --verbose
GET /health HTTP/1.1
User-Agent: check_http/v2.3.1 (nagios-plugins 2.3.1)
Connection: close
Host: localhost:8428
Accept: */*


CRITICAL - Socket timeout

Every now and then, I do get successful return of OK:

/usr/lib64/nagios/plugins/check_http -H localhost -u /health -p 8428 --verbose
GET /health HTTP/1.1
User-Agent: check_http/v2.1.4 (nagios-plugins 2.1.4)
Connection: close
Host: localhost:8428
Accept: */*


http://localhost:8428/health is 122 characters
STATUS: HTTP/1.1 200 OK
**** HEADER ****
Content-Type: text/plain
Date: Fri, 27 Dec 2019 14:37:08 GMT
Content-Length: 2
Connection: close

**** CONTENT ****
OK
HTTP OK: HTTP/1.1 200 OK - 122 bytes in 0.002 second response time |time=0.001788s;;;0.000000 size=122B;;;0

curl returns successful result every time (notice, there is no new line at the end of OK string):

curl -v http://localhost:8428/health
* About to connect() to localhost port 8428 (#0)
*   Trying ::1...
* Connection refused
*   Trying 127.0.0.1...
* Connected to localhost (127.0.0.1) port 8428 (#0)
> GET /health HTTP/1.1
> User-Agent: curl/7.29.0
> Host: localhost:8428
> Accept: */*
>
< HTTP/1.1 200 OK
< Content-Type: text/plain
< Date: Fri, 27 Dec 2019 14:40:05 GMT
< Content-Length: 2
<
* Connection #0 to host localhost left intact
OK

I'm not sure if this is a bug on check_http or VM side, since curl is returning OK everytime. If I run check_http with -N flag, then it passes the check.

[sawolf]

Hi @matejzero, thanks for reporting this. I'm not sure when I'll be able to debug/investigate this, but at least I was able to reproduce this against their container. You said it works occasionally, so I suspect this is a bug in the nagios-plugins project.

[matejzero]

Sure, no problem. It works for me with -N flag for now, so I'm fine. I don't need string check for now, since status code is enough in my case.

[sawolf]

The pull request above seems to fix the issue in my environment. Would you be willing to compile and verify that it works for you as well?

[matejzero]

I can confirm this PR fixes the bug.

# ./check_http -H localhost -u /health -p 8428
HTTP OK: HTTP/1.1 200 OK - 122 bytes in 0.002 second response time |time=0.001807s;;;0.000000 size=122B;;;0