Failed to fetch nginx stub status page

[EHRETic]

Hi there,

I'm new to Zabbix (started a week ago with the 5.2) and it's my very first post here.

I'm starting installing slowly but surelly Zabbix agent on all my VMs. It's in my lab so there is no panic, but I always wanted an overview of what is going on in my infra as it's start to be quite large for a homelab...
I've a few VMs that use NGINX as a https web proxy for different services that use a web console (Zabbix is one of them BTW), I find NGINX quite convenient for that. Of course I'd like to monitor it.

So I've attached the template "Nginx by Zabbix agent" to the necessary hosts and straight away, I got "Service is down" alerts.
I've solved this one by replacing the key in some of the items (solution was provided here: https://www.zabbix.com/forum/zabbix-...ervice-is-down) but the despite that change, the alert "Nginx: Failed to fetch stub status page (or no data for 30m)" is still popping up. Also, I don't know if it is "right" to change it as the problem might have another source.

The changes are in yellow, changing the red one didn't help :



Issue is the same if I use "Nginx by HTTP" template.

Any idea where I should start? Thanks in advance

[EHRETic]

Hi there,

After a lot of other things to do, I was today going back at that one. So I've tried all mentioned above (and my server is also migrated to 6.0 LTS), but it didn't work.

So I looked further and I found out that you HAVE TO publish at some point the NGINX status page in order to read (sorry, but I had to learn that).
In my case, I choose to create a dedicated server listening on 127.0.0.1 so I would not disturb any running configuration.

So I added the following configuration block on my NGINX servers and reconfigured the template accordingly, {$NGINX.STUB_STATUS.PATH} remained empty.

NGINX server block:

# NGINX monitoring
server {
    listen 8080;
    server_name 127.0.0.1;

    location / {
        stub_status on;
        allow 127.0.0.1;
        deny all;
    }
}

Thanks all for you help anyway, I hope this will serve to others!

[dereitz]

Were you able to resolve this?

I'm also new to Zabbix. I'm running both the Zabbix server and the agent in Docker containers which has added to the complexity.

In any case, I cannot seem to get either the NGINX via HTTP or the NGINX via Zabbix agent templates working while inside the Docker container. Has anyone had success with this? I'd appreciate any tips!!!

[cyber]

"Nginx by Zabbix agent" does provide these macros... So that info in that other page you linked, is wrong... no need to change anything in keys.
{$NGINX.STUB_STATUS.HOST}","{$NGINX.STUB_STATUS.PA TH}","{$NGINX.STUB_STATUS.PORT}" default values in template are localhost, basic_status and 80. So agent does locally a web.get from same host. Maybe your nginx is not configured to provide that page to localhost? Can you get it on command line with curl or wget?

"Nginx by HTTP" solves same with http call to {$NGINX.STUB_STATUS.SCHEME}://{HOST.CONN}:{$NGINX.STUB_STATUS.PORT}/{$NGINX.STUB_STATUS.PATH} basically http://<your host>:80/basic_status by default. Is there something answering on that port, is the port reachable from server/proxy? FW-s open? Again, you can test from server/proxy with wget/curl...

[tariver]

Sorry for necroposting, but this is the first result that comes on Google for "Zabbix Nginx Service status down" query.

Have the same problem, turned out that Zabbix agent uses IPv6 to connect to nginx and my nginx were not using IPv6.
To check if this item works locally:

zabbix_agentd -t net.tcp.service[http,localhost,80]

The output should be

net.tcp.service[http,localhost,80]            [u|1]

If the last part is [u|0] then something is wrong.

To check if nginx is listening on IPv6

$ sudo netstat -tnlp | grep nginx

There should be a line that reads something to this:

tcp        0      0 :::80                       :::*                        LISTEN      7815/nginx

If there is no such a line then your nginx doesn't.

How to fix it: add the following line to server configuration where the basic_statusis is defined.

listen [::]:80;

or this line if you want nginx to listen to IPv6 only on localhost

listen [::1]:80;

After that check your nginx configuration with

nginx -t

and then reload config (doesn't affect users)

nginx -s reload

[RonaldBag]

I found a simpler solution: Change the macro to use 127.0.0.1 instead of localhost.

[EHRETic]

Hi there,

After a lot of other things to do, I was today going back at that one. So I've tried all mentioned above (and my server is also migrated to 6.0 LTS), but it didn't work.

So I looked further and I found out that you HAVE TO publish at some point the NGINX status page in order to read (sorry, but I had to learn that).
In my case, I choose to create a dedicated server listening on 127.0.0.1 so I would not disturb any running configuration.

So I added the following configuration block on my NGINX servers and reconfigured the template accordingly, {$NGINX.STUB_STATUS.PATH} remained empty.

NGINX server block:

# NGINX monitoring
server {
    listen 8080;
    server_name 127.0.0.1;

    location / {
        stub_status on;
        allow 127.0.0.1;
        deny all;
    }
}

Thanks all for you help anyway, I hope this will serve to others!