Metrics not being exposed even though all dependencies installed and enabled

[janholbrouck]

I have been trying to add your exporter to my Ubuntu Xenian machine running RMQ. I am running RabbitMQ v3.6.6 so have installed and enabled all the dependencies found here: https://github.com/deadtrickster/prometheus_rabbitmq_exporter/releases/tag/rabbitmq-3.6.14.1

I ran 'sudo rabbitmq-plugins list', displaying all the enabled plugins, and all the dependencies are enabled, so I would expect it to work.

I have not set a specific endpoint in my rabbitmq.config file, even though I have tried without result.

The problem I have is that no matter which URL I surf to, I can not seem to get the exported metrics.

The URL I use for this is UbuntuMachineIPadress:15672/api/metrics I can perfectly access my RMQ management console with UbuntuMachineIPAdress:15672/#/ however. I've been trying for a while to figure out why this wouldn't be working, but I can't really figure out why.

When I installed docker on my Ubuntu machine and ran the command "docker run -p 8080:15672 deadtrickster/rabbitmq_prometheus" I managed to get exposed metrics, but not for the local RMQ I am running on the machine.

Any help would be appreciated!

[deadtrickster]

hi, what your browser says when you open .../api/metrics? are you seeing something in the rabbitmq logs?

[janholbrouck]

Hi. thanks for the quick response. My browser says this: "{"error":"Object Not Found","reason":"\"Not Found\"\n"}"

I have stumbled on what might be the problem though, and am trying some more things, Since I am running RabbitMQ 3.6.6 I am unsure which release I should be using. I initially used 3.6.x since I figured that referred to 3.6.6 aswel, but I am now thinking I should perhaps try one of the earlier releases, although you don't seem to have one for 3.6.6 exactly.

My logs did not display any odd information

[janholbrouck]

So I have not yet got it to work with on of the other 2 releases (3.6.5 and 3.6.8/9), I am getting the error {plugin_module_unloadable,"prometheus_process_collector", {error,on_load_failure}}

I have tried following the steps from https://github.com/deadtrickster/prometheus_rabbitmq_exporter/issues/26 but couldn't really figure it out. I am able to remove the folder, and also able to create the folder as the rabbitmq user, so I am not sure if this is the same problem that jperville described or not.

I added a system variable in /etc/environment RABBITMQ_PLUGINS_EXPAND_DIR="/tmp/rabbitmq-expand-plugins"

When I start the rabbitmq service this folder is also created, so all seemed okay to me. I am however not sure if the /tmp folder has a noexec set or not, and I am unsure on how to check this.

[deadtrickster]

to check noexec run mount. also, have you tried without process collector?

---

From: janholb notifications@github.com Sent: Wednesday, October 31, 2018 5:23:40 PM To: deadtrickster/prometheus_rabbitmq_exporter Cc: Ilya Khaprov; Comment Subject: Re: [deadtrickster/prometheus_rabbitmq_exporter] Metrics not being exposed even though all dependencies installed and enabled (#61)

So I have not yet got it to work with on of the other 2 releases (3.6.5 and 3.6.8/9), I am getting the error {plugin_module_unloadable,"prometheus_process_collector", {error,on_load_failure}}

I have tried following the steps from #26https://github.com/deadtrickster/prometheus_rabbitmq_exporter/issues/26 but couldn't really figure it out. I am able to remove the folder, and also able to create the folder as the rabbitmq user, so I am not sure if this is the same problem that jperville described or not.

I added a system variable in /etc/environment RABBITMQ_PLUGINS_EXPAND_DIR="/tmp/rabbitmq-expand-plugins"

When I start the rabbitmq service this folder is also created, so all seemed okay to me. I am however not sure if the /tmp folder has a noexec set or not, and I am unsure on how to check this.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/deadtrickster/prometheus_rabbitmq_exporter/issues/61#issuecomment-434750034, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AAKhF9gno13-lYHY0KMd8Qe5wpcW3A6wks5uqc6MgaJpZM4YEPvJ.

[janholbrouck]

I have tried without the process collector, then everything runs just fine but I still don't get any metrics. Browsing to /api/metrics gives me this: {"error":"Object Not Found","reason":"\"Not Found\"\n"}

Running the "mount" command gives me this:

I added a system variable in /etc/environment RABBITMQ_PLUGINS_EXPAND_DIR="/tmp/rabbitmq-expand-plugins" I don't really see anything about the /tmp folder in the mount command though. then again, /tmp is a folder so can't show up there itself, can it?

[janholbrouck]

I have gotten it to work by trying a different release (3.6.5) and disabling the prometheus_process_collector. It was confusing to me, since my RabbitMQ version is 3.6.6. Thanks for the help and project!