Is your feature request related to a problem? Please describe.

The current automated releases are only for x86 architecture.

Describe the solution you'd like
Updating the current CI to build for multiple architectures (ie docker buildx)

Describe the bug
Most recent image on DockerHub (0.5.0, ef5810d4ce30) is based on outdated alpine image with security vulnerabilities.

To reproduce

$ docker pull nginx/nginx-prometheus-exporter:0.5.0
$ docker run -it --entrypoint=/bin/sh nginx/nginx-prometheus-exporter:0.5.0
$ cat /etc/alpine-release 
3.9.4
$ apk upgrade -s
(1/4) Upgrading musl (1.1.20-r4 -> 1.1.20-r5)
(2/4) Upgrading libcrypto1.1 (1.1.1b-r1 -> 1.1.1d-r2)
(3/4) Upgrading libssl1.1 (1.1.1b-r1 -> 1.1.1d-r2)
(4/4) Upgrading musl-utils (1.1.20-r4 -> 1.1.20-r5)
OK: 6 MiB in 14 packages

Base image is from June 2019...

Expected behavior
Image should be based on a recent base image that does not contain any security vulnerabilities.

Your environment
Docker image 0.5.0 - ef5810d4ce30

Additional context
Perhaps one can clarify, if this is intended to be a production grade image or just for development purposes only.

Describe the bug
When I set CONST_LABELS, it does not make any effect . When I try to set -prometheus.const-labels, it produces an error.

To reproduce
Steps to reproduce the behavior:

1. docker run -e CONST_LABELS=blabla=blabla -it nginx/nginx-prometheus-exporter:0.5.0 -nginx.scrape-uri http://hostname/basic_status
2. Go to container_ip:9113/metrics
3. See metrics without label(s) passed before

OR

1. docker run -it nginx/nginx-prometheus-exporter:0.5.0 -nginx.scrape-uri http://hostname/basic_status -prometheus.const-labels blabla=blabla
2. See error: flag provided but not defined: -prometheus.const-labels

Expected behavior
Metrics are shown with labels.

Your environment

• Version of the Prometheus exporter: gitCommit="c3dd65c",version="0.5.0"
• Version of Docker/Kubernetes: Docker version 18.09.6, build 481bc77
• Using NGINX

Hello.

Just took a quick look on it and deployed over my hosts. There's a few things that might add to the value:

1. It would be nice to have _build_info. Currently I can't get exporter info on the board to plan tiered upgrades.

2. It would be nice to be able to scrape multiple hosts. With HTTPS/Auth nginx is already an external service that can be used to transfer data. This will also make it possible to transport data securely over network. For example, if I have a few hosts in different DCs I need to proxy exporter via nginx for example, if exporter would support scraping more then one host, authorization and HTTPS it would be possible to expose /stub_status on nginx externally in a restricted location. This will make provisioning easier.

Prerequisites
We assume that you have already installed Prometheus and NGINX or NGINX Plus. Additionally, you need to:

Expose the built-in metrics in NGINX/NGINX Plus:
For NGINX, expose the stub_status page at /stub_status on port 8080.

Running the Exporter in a Container
To start the exporter we use the docker run command.

To export NGINX metrics, run:

$ docker run -p 9113:9113 nginx/nginx-prometheus-exporter:0.1.0 -nginx.scrape-uri http://:8080**/api**

Shouldn't it be $ docker run -p 9113:9113 nginx/nginx-prometheus-exporter:0.1.0 -nginx.scrape-uri http://:8080**/stub_status** ?

I have a nginx container running on localhost:70/nginx_status
I have a prometheus container running on localhost:9090

when i run
docker run -it -p 9113:9113 --network bridge nginx/nginx-prometheus-exporter:0.1.0 -nginx.scrape-uri http://localhost:70/nginx_status

error
Could not create Nginx Client: Failed to create NginxClient: failed to get http://localhost:70/nginx_status: Get http://localhost:70/nginx_status: dial tcp 127.0.0.1:70: connect: connection refused

docker ps -a shows that nginx-prometheus(stoic_goldberg) exited

how do i get this container to be on the same default network(bridge)
the connection failed error is because it does not have access to http://localhost:70/nginx_status because it is not on the same "bridge" network

Hi,

I would like to use your nginx-prometheus-exporter as a sidecar container in kubernetes to collect metrics from a nginx container running in the same Pod.
Unfortunately the nginx container needs ~1-2 seconds to start. So the scrape target for the nginx-prometheus-exporter is not reachable when it starts. Therefore the exporter stops and the container has to be restarted from kubernetes. After the container restart, the exporter is working fine because the nginx process was able to start in the meantime and the scrape target is reachable then.

So I would like to ask if it would be possible to add a parameter to the exporter to avoid the default behavior of stopping. A simple sleep and retry after a specific time (maybe 5s) would be very useful.

The related error log output:

root@k8s [dc]:~# kubectl logs $pod -c prometheus-exporter  -p

2019/02/20 11:42:51 Starting NGINX Prometheus Exporter Version=0.3.0 GitCommit=6570275
2019/02/20 11:42:51 Could not create Nginx Client: Failed to create NginxClient: failed to get http://127.0.0.1:8080/stub_status: Get http://127.0.0.1:8080/stub_status: dial tcp 127.0.0.1:8080: connect: connection refused

Thank you.

I am trying to install nginx-prometheus-exporter on my system. While applying the "make" command in my $GOPATH/src/github.com/nginxinc/nginx-prometheus-exporter directory, I get an error stating.

flag provided but not defined: -mod

What could be the possible error of that? Thank you

nginx-prometheus-exporter -nginx.scrape-uri http://127.0.0.1:6666/stub_status

2020/09/07 02:13:14 Starting NGINX Prometheus Exporter Version=0.8.0 GitCommit=de15093
2020/09/07 02:13:14 Could not create Nginx Client: failed to parse response body "Active connections: 1 \nserver accepts handled requests request_time\n 55 55 55 175255\nReading: 0 Writing: 1 Waiting: 0 \n": invalid input for connections and requests " 55 55 55 175255"

When using the nginxplus_upstream_server_state it is frustrating, that the different states are integers. This makes it hard to do a count(nginxplus_upstream_server_state) group by (upstream, state) query(e.g. for multiple loadbalancers) in prometheus.

Describe the solution you'd like
It is far more common in prometheus, to have different timeseries for different statuses as long as they have an upper bound. This is also the way upstream error codes are currently handled.

Additional context
upstreamServerStates mapping:

How status codes are handled:

Is there a reason such a mapping to integers was chosen?

How can I get http errors? 2xx 4xx 5xx

Thank you

Is your feature request related to a problem? Please describe.
Using nginx-prometheus-exporter in security-enhanced (like active PodSecurityPolicy) Kubernetes (or OpenShift) cluster requires non-root containers. It is common to use scratch image to reduce attack surface and get a smaller final image.

Describe the solution you'd like

• Non-root container - USER in Dockerfile
• Use scratch image

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
nginx-prometheus-exporter is written in Go, thus alpine:3.11 is not needed to run it.

Describe the bug
When running the docker container locally pointing it against our nginx lb api endpoint exposed on port 8888 I can see that errors have occurred when gathering the metrics:

An error has occurred during metrics gathering:
490 error(s) occurred:
* collected metric nginxplus_upstream_server_state label:<name:"server" value:"127.0.0.1:80" > label:<name:"upstream" value:"gru_location_red" > gauge:<value:3 >  was collected before with the same name and label values
* collected metric nginxplus_upstream_server_active label:<name:"server" value:"127.0.0.1:80" > label:<name:"upstream" value:"gru_location_red" > gauge:<value:0 >  was collected before with the same name and label values
* collected metric nginxplus_upstream_server_requests label:<name:"server" value:"127.0.0.1:80" > label:<name:"upstream" value:"gru_location_red" > counter:<value:0 >  was collected before with the same name and label values
* collected metric nginxplus_upstream_server_responses label:<name:"code" value:"1xx" > label:<name:"server" value:"127.0.0.1:80" > label:<name:"upstream" value:"gru_location_red" > counter:<value:0 >  was collected before with the same name and label values
* collected metric nginxplus_upstream_server_responses label:<name:"code" value:"2xx" > label:<name:"server" value:"127.0.0.1:80" > label:<name:"upstream" value:"gru_location_red" > counter:<value:0 >  was collected before with the same name and label values
* collected metric nginxplus_upstream_server_responses label:<name:"code" value:"3xx" > label:<name:"server" value:"127.0.0.1:80" > label:<name:"upstream" value:"gru_location_red" > counter:<value:0 >  was collected before with the same name and label values
* collected metric nginxplus_upstream_server_responses label:<name:"code" value:"4xx" > label:<name:"server" value:"127.0.0.1:80" > label:<name:"upstream" value:"gru_location_red" > counter:<value:0 >  was collected before with the same name and label values
* collected metric nginxplus_upstream_server_responses label:<name:"code" value:"5xx" > label:<name:"server" value:"127.0.0.1:80" > label:<name:"upstream" value:"gru_location_red" > counter:<value:0 >  was collected before with the same name and label values
...
...
...
And the list goes on and on.

Our Prometheus can't gather the metrics because of these errors.

To reproduce
Steps to reproduce the behavior:

1. Deploy using docker run -p 9113:9113 nginx/nginx-prometheus-exporter:0.1.0 -nginx.plus -nginx.scrape-uri http://**lb_address**:8888/api
2. View http://localhost:9113/metrics
3. Observe the errors

Expected behavior
Metrics should be gathered successfully, allowing us and prometheus to pick these up.
If it's something we're doing wrong from the lb side it should be mentioned.

Your environment

• Version of the Prometheus exporter - 0.1.0
• Version of Docker/Kubernetes - Docker version 18.06.1-ce, build e68fc7a
• [if applicable] Kubernetes platform (e.g. Mini-kube or GCP)
• Using NGINX or NGINX Plus - NGINX Plus

Additional context
Add any other context about the problem here. Any log files you want to share.

Can you change the configuration instead of port 8080 that listens on port 8888 ???

Is your feature request related to a problem? Please describe.
I want to monitor multiple nginx instances across a fleet of services with the same prometheus instance. This doesn't work well since each instance of this exporter uses the same metric names and will clobber each other. I'd like to add support for labels in the metric so we can annotate the name of the service to support grouping by those.
I'm a novice prometheus user so let me know if this is already somehow possible.

Describe the solution you'd like
Allow the exported metrics to have labels. Eg: nginx_connections_handled{name="app1"} 2918

Describe alternatives you've considered
I cannot think of another way. Is it possible to change the name or prefix the exported metrics? That would work as well but I think labels are probably nicer.

Additional context
N/A

Hi,
in your actually version of the exporter, the result of "curl -s 127.0.0.1/api/3/stream/zone_sync" is missing. This is a cluster configuration part in nginx+ which ist activated for sticky learn to sync all cookie in an active/active cluster. Could you please add this values?

e.g. of result:

{ "status": { "nodes_online": 1, "msgs_in": 316063, "msgs_out": 312921, "bytes_in": 619285360, "bytes_out": 616452834 }, "zones": { "example_zone_one": { "records_total": 153746, "records_pending": 11 }, "example_zone_two": { "records_total": 10940, "records_pending": 18 } } }

Describe the bug
Metric with name nginxexporter_build_info is empty while at 0.4.2 it contains correct info.

To reproduce
Steps to reproduce the behavior:

1. Deploy using 0.6.0 version
2. Check metrics
3. Example output:

# HELP nginxexporter_build_info Exporter build information
# TYPE nginxexporter_build_info gauge
nginxexporter_build_info{gitCommit="",version=""} 1

Expected behavior
Version and commit should appear

Your environment

• Version of the Prometheus exporter - 0.6.0
• Version of Docker/Kubernetes: 1.15.11-gke1
• Kubernetes platform: GCP
• NGINX 1.16.1-1

Hi, any chance of creating a release with the recent changes ?
I would love to be able to use a newer version in my environment.
@Rulox

Hello,

on one node_exporter can add several parameters -nginx.scrape-uri?

for example

-nginx.scrape-uri http://localhost:8081/nginx_status http://localhost:8082/nginx_status

Would it be possible to mimic how the blackbox exporter behaves and allow probing multiple nginx servers? I'd like to avoid running one nginx-prometheus-exporter per nginx server when we have hundreds of nginx servers. The official blackbox exporter allows you to make an API call to it with the target in the URL. For example:

http://nginx-prometheus-exporter-container:9115/probe?target=nginx-server01.domain.local
http://nginx-prometheus-exporter-container:9115/probe?target=nginx-server02.domain.local

Thanks for the consideration!

Since NGINX Plus R18 the API module provides the list of fields according to the configuration instead of using the static list of fields.

In previous versions, the api returned the static list of fields: "nginx","processes","connections","slabs","http","stream","ssl"

Since r18, this list can be different for diffrent configurations. For example, if there are no configured stream{}, then the list of fields will be:
"nginx","processes","connections","slabs","http""ssl"

To reproduce
Steps to reproduce the behavior:

1. Upgrade NGINX Plus up to the latest r18 release
2. Comment out the whole stream {} configuration
3. See error

Expected behavior
The prometheus exporter should request /api/ first, to get the number of latest available API version. After that, request /api/ to get the list of configured fields.

I have a little swarm cluster with 6 nodes and I deploy 2 replicas of nginx on it:

nginx:
  image: registry:5000/nginx:2
  ports:
    - "80:80"
    - "443:443"
  deploy:
    replicas: 2

Is it possible to scape metrics from every replica via nginx-exporter ?

Hi,

could you please add a value which has timestamp last reload Nginx? like:

curl -s '127.0.0.1/api/4/nginx' | jq
{
"version": "1.15.10",
"build": "nginx-plus-r18",
"address": "127.0.0.1",
"generation": 7,
"load_timestamp": "2019-05-30T11:15:03.206Z",
"timestamp": "2019-05-30T11:44:16.317Z",
"pid": 60024,
"ppid": 31797
}

Also I can use the rest of values like version, build etc.

Tanks a lot

Adding nginx-prometheus-exporter to the distro repos at http://nginx.org/packages/ would be useful.

Hi,

• Is it possible to introduce a health check endpoint for the exporter itself?
• Is it possible to introduce specifying the configuration parameters as a file, in addition to command line flags?

Thanks!

i'm trying to use binary but unfortunately failed. can you please describe how to configure for arm64 base os.

We are already using your exporter but unfortunately there is one thing we are missing to migrate from our zabbix based nginx plus monitoring to grafana / prometheus. We need somehow the number of upstreams that have been discovered and those who are up and down.

As my knowlege of the Go Language is limited i cannot contribute it myself.

Regards,

Mike

hi I trued to use nginx-prometheus-exporter with tengine (fork of ngixn) and faced with problem
after start of the exporter i see logs like

2020/08/17 13:53:44 Starting NGINX Prometheus Exporter Version= GitCommit=
2020/08/17 13:53:44 Could not create Nginx Client: failed to parse response body "Active connections: 4 \nserver accepts handled requests request_time\n 9055 9055 65927 271898450\nReading: 0 Writing: 2 Waiting: 2 \n": invalid input for connections and requests " 9055 9055 65927 271898450"

is it possible to work with the web server? or only support official versions of nginx ?

Can you add metrics for nginx cache? I see that metrics already exists via "/api/4/http/caches" nginx url.

Describe the bug
Not sure this is a bug or intentionally omitted, but const labels are not attached to the upMetric while it is present on all other metrics.
For the use case of my team we want to show, in Grafana, the number of pods that are up and running and also the version of our build being served by the Nginx in those pods. The number of pods we get by just summing the upMetric but we don't get the version because that is not included in the upMetric

To reproduce
Steps to reproduce the behavior:

1. Run the nginx-prometheus-exporter docker container with CONST_LABELS env. variable set to version=1.2.1
2. Visit the /metrics url to get a list of all the metrics
3. Compare one of the stub_stats metrics with the upMetric and verify that version=1.2.1 is returned by the stub_stats metrics but not the upMetric

Expected behavior
I'd like the upMetric to include const labels just lite the stub_stats metrics

Your environment

• Version of the Prometheus exporter: 0.6.0 release
• Version of Kubernetes: v1.11.0+d4cacc0
• [if applicable] Kubernetes platform: OpenShift v3.11.135
• Using NGINX

Additional context

Hello.

I was playing with different exporters and noticed that this exporter is quitting when I'm pkill -1 prometheus to renew configuration. It's not dying, I see no cores or messages about that, it just exits when connection from prometheus is closed.

Describe the bug
Every release note contains the wrong compatibility version at https://github.com/nginxinc/nginx-prometheus-exporter/releases

COMPATIBILITY:
NGINX 0.1.18 or newer.

While 0.16 is the last stable

To reproduce
Steps to reproduce the behavior:

1. Open releases page https://github.com/nginxinc/nginx-prometheus-exporter/releases
2. Check any release

Expected behavior
Release notes contains correct compatibility version

Describe the bug
error when starting the exporter I get the following error:

response: json: cannot unmarshal object into Go value of type client.versions

To reproduce
Steps to reproduce the behavior:

Run ./nginx-prometheus-exporter -nginx.plus -nginx.scrape-uri http://localhost/status/ locally to my nginx instance.

Running nginx 1.2.2

Expected exporter to run

Your environment
NGINX Plus V1.2.2

Is your feature request related to a problem? Please describe.
Many teams want to define a latency SLO for their HTTP services. An average request latency is not the best choice, because it is not possible to know if there was a certain number of requests that took too long.

Describe the solution you'd like
Nginx exporter should expose a response latency bucket metric, like described in the prometheus documentation: https://prometheus.io/docs/practices/histograms/#quantiles

Describe alternatives you've considered
There is no alternative, except for using a different exporter or calculating the response time buckets in some other way. The average is not enough for calculation of percentiles.

Additional context
https://landing.google.com/sre/sre-book/chapters/service-level-objectives/

Do you know when the next release will be? I'm particularly interested in a Docker image containing #103 being released

Thanks!

I have one access.log for all virtual hosts on the machine. With a log format like this:

    log_format main '$remote_addr - $remote_user [$time_local] '
            '"$host" "$request" $status $body_bytes_sent '
            '$request_time $upstream_response_time '
            '"$http_referer" "$http_user_agent"';

Is there any way I can filter metrics just for one host?
Or can I group metrics by host name?

Thanks!

Hi,

I'm trying to run nginx_exporter with few flags and I'm still getting this message

flag provided but not defined: -nginx.ssl-ca-cert

here is how I run it
/usr/local/bin/nginx_exporter -nginx.scrape-uri "https://127.0.0.1:443/status" -nginx.ssl-ca-cert "/etc/pki/tls/cert.pem"
I tried without quotations and get the same results.

I'm using nginx_exporter versions 0.6.0

Any ideas ?

Exporter was throwing some errors and metrics are missing at the same time.
It happened randomly and recovered later. Here is logs

-- Logs begin at Tue 2020-12-01 04:59:14 UTC. --
Dec 02 00:21:59 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:21:59 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Dec 02 00:22:04 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:22:04 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Dec 02 00:22:09 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:22:09 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": context deadline exceeded
Dec 02 00:22:14 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:22:14 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Dec 02 00:22:19 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:22:19 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Dec 02 00:22:24 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:22:24 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Dec 02 00:22:59 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:22:59 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Dec 02 00:23:19 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:23:19 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": context deadline exceeded (Client.Timeout exceeded while awaiting headers)
Dec 02 00:23:29 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:23:29 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Dec 02 00:27:19 instance-name nginx-prometheus-exporter[3550]: 2020/12/02 00:27:19 Error getting stats: failed to get http://127.0.0.1:8080/stub_status: Get "http://127.0.0.1:8080/stub_status": context deadline exceeded (Client.Timeout exceeded while awaiting headers)

Environment

• Version of the Prometheus exporter

 Starting NGINX Prometheus Exporter Version=0.8.0 GitCommit=de15093

• Version of VM

Distributor ID: Ubuntu
Description:    Ubuntu 16.04.3 LTS
Release:        16.04
Codename:       xenial

• Using NGINX or NGINX Plus

nginx version: nginx/1.10.3 (Ubuntu)

There is no nginx logs as well in the same time

root@instance-name:/home/thamaraiselvam# journalctl -fu nginx
-- Logs begin at Tue 2020-12-01 04:59:14 UTC. --
Dec 01 05:19:21 instance-name systemd[1]: Stopping A high performance web server and a reverse proxy server...
Dec 01 05:19:21 instance-name systemd[1]: Stopped A high performance web server and a reverse proxy server.
Dec 01 05:19:21 instance-name systemd[1]: Starting A high performance web server and a reverse proxy server...
Dec 01 05:19:21 instance-name systemd[1]: nginx.service: Failed to read PID from file /run/nginx.pid: Invalid argument
Dec 01 05:19:21 instance-name systemd[1]: Started A high performance web server and a reverse proxy server.
Dec 01 05:32:08 instance-name systemd[1]: Stopping A high performance web server and a reverse proxy server...
Dec 01 05:32:08 instance-name systemd[1]: Stopped A high performance web server and a reverse proxy server.
Dec 01 05:32:08 instance-name systemd[1]: Starting A high performance web server and a reverse proxy server...
Dec 01 05:32:09 instance-name systemd[1]: nginx.service: Failed to read PID from file /run/nginx.pid: Invalid argument
Dec 01 05:32:09 instance-name systemd[1]: Started A high performance web server and a reverse proxy server.

Is your feature request related to a problem? Please describe.
I wouldn't call it a problem, but there is no _up metric exported.

Describe the solution you'd like
I would like to have _up metric exported so it's easier to write an alert that NGINX is down.

Describe alternatives you've considered
None yet.

Additional context
Maybe there is another way to write the above alert with the data that's already exported but I am not aware of it yet.

Thank you.

When running

the README states

http://127.0.0.1:8080/stub_status is the default scrape URI. Running it without changing it seems to confirm.

2018/08/11 00:12:14 Could not create Nginx Client: Failed to create NginxClient: failed to get http://127.0.0.1:8080/stub_status: Get http://127.0.0.1:8080/stub_status: dial tcp 127.0.0.1:8080: connect: connection refused

However, if I try to define it as that via -nginx.scrape-uri="http://127.0.0.1:8080/stub_status", when I run it I get:

2018/08/11 00:11:20 Could not create Nginx Client: Failed to create NginxClient: failed to get "http://127.0.0.1:8080/stub_status": parse "http://127.0.0.1:8080/stub_status": first path segment in URL cannot contain colon

The output is as follows

2020/06/12 13:27:32 Starting NGINX Prometheus Exporter Version=0.7.0 GitCommit=a2910f1
2020/06/12 13:27:32 Could not create Nginx Client: failed to get http://myyyy:80/nginx_status: Get http://myyyy:80/nginx_status: dial tcp: lookup myyyy on 172.21.0.10:53: no such host

”myyyy“ is a service in my k8s
but it don't work！！！

Hi.

Describe the bug
There is no argument nginx.ssl-verify

To reproduce
Steps to reproduce the behavior:

1. docker run -p 9113:9113 nginx/nginx-prometheus-exporter:0.1.0 -nginx.scrape-uri https://site.com/basic_status
   2018/10/05 16:30:56 Starting NGINX Prometheus Exporter Version=0.1.0 GitCommit=8d90a86
   2018/10/05 16:30:56 Could not create Nginx Client: Failed to create NginxClient: failed to get https://site.com/basic_status: Get https://site.com/basic_status: x509: failed to load system roots and no roots provided
2. docker run -p 9113:9113 nginx/nginx-prometheus-exporter:0.1.0 -nginx.scrape-uri https://site.com/basic_status -nginx.ssl-verify false
   flag provided but not defined: -nginx.ssl-verify
   Usage of /usr/bin/exporter:
   -nginx.plus
   Start the exporter for NGINX Plus. By default, the exporter is started for NGINX.
   -nginx.scrape-uri string
   A URI for scraping NGINX or NGINX Plus metrics.
   For NGINX, the stub_status page must be available through the URI. For NGINX Plus -- the API. (default "http://127.0.0.1:8080/stub_status")
   -web.listen-address string
   An address to listen on for web interface and telemetry. (default ":9113")
   -web.telemetry-path string
   A path under which to expose metrics. (default "/metrics")

I can see Nginx throws 2 types of Monitoring Logs:

• Stub Status
• Prometheus Metrics

Only Stub Status is being considered here. I can setup 2 services to get those in 1 Prometheus Server, but I still have some following doutbts:

• Why Prometheus Metrics are not considered ?
• Why Nginx is throwing 2 different metrics ?
• How to combine them together ?
• Are there any future plan to include Prometheus Metrics in this too ?

After importing dashboard all panels of the dashboard is filled with "Test data: random walk" and Query is "default" not Prometheus as selected during import. Some more details can be found here.

Steps to reproduce the behavior:

1. Import dashboard with one of the latest version of Grafana (I'm using 6.4.3)
2. See all panels are filled with random data.

Expected behavior
Panels displaying actual data, datasource is correctly set.

Environment

• NGINX Prometheus Exporter Version=0.4.2 GitCommit=f017367
• Grafana 6.4.3
• NGINX

Additional context
I suppose problem can be fixed by re-exporting dashboard with Grafana 6.4+. Or by manually adding "datasource": "${DS_PROMETHEUS}", into each panel in JSON data, this is how I fixed it for myself.

how to scrape from multiple domains/servers?

i tried args like that

- "--nginx.scrape-uri=https://domain1.kz/basic_status, https://domain2.kz/basic_status"

but prometheus showing only one.

Describe the bug
im running nginx exporter using docker compose
i have this compose file

`
version: '2'
services:

prometheus:
image: nginx/nginx-prometheus-exporter:0.2.0
user: root
volumes:
- /monitoring:/monitoring
command:
- '-nginx.scrape-uri http://x.x.x.x:80/nginx_status'
ports:
- "9113:9113"
restart: unless-stopped

`
when running this command docker-compose up i got this error

flag provided but not defined: -nginx.scrape-uri http://x.x.x.x:80/nginx_status

Describe the bug
Nginx exporter fails to export metrics

Exporter was failing on 0.3.0 to export metrics with path not found error after upgrade to 0.4.2. Some of the hosts are able to export metrics. But there is one host still has same error.

Expected behavior
Export metrics without failure.

Your environment
Centos 3.10.0-957.1.3.el7.x86_64

nginx version: nginx/1.15.10 (nginx-plus-r18-p1)
built by gcc 4.8.5 20150623 (Red Hat 4.8.5-36) (GCC)
built with OpenSSL 1.0.2k-fips 26 Jan 2017

nginx exporter version 0.4.2

Additional context

# Systemd command
nginx-prometheus-exporter -web.listen-address :9113 -nginx.plus -nginx.scrape-uri http://127.0.0.1:8888/api -nginx.ssl-verify false

# Error log
nginx-prometheus-exporter[13565]: 2019/09/26 21:17:14 Error getting stats: failed to get stats: failed to get stream server zones: expected 200 response, got 404. path=; method=; error.status=404; error.text=path not found; error.code=PathNotFound; request_id=10e67745e5d4466f570cb834aa9c17e0; href=https://nginx.org/en/docs/http/ngx_http_api_module.html

Describe the solution you'd like
It is worthwhile to know the number/graph of dropped connections for Nginx e.g. nginx_connections_dropped

Describe alternatives you've considered
This can be retrieved from (accepted-handled) with PromQL in Prometheus/Grafana.

I've been looking for a grafana dashboard at https://grafana.com/dashboards
but cannot find any which would mentioned compatibility with this exporter.

Is there any dashboard available? If yes, please mention in the readme.

Thank you very much.

Hello,

Could you help me with the problem in using docker for EC2 (AWS) instances?
It happens that since I have instances configured in auto-scalling, for me it is much easier if the initialization string accepts something like localhost and / or 127.0.0.1.
But when I try to boot using them the following failures occur:

docker run -p 9113:9113 nginx/nginx-prometheus-exporter:0.2.0
2018/12/05 20:24:51 Starting NGINX Prometheus Exporter Version=0.2.0 GitCommit=ad0a472
2018/12/05 20:24:51 Could not create Nginx Client: Failed to create NginxClient: failed to get http://127.0.0.1:8080/stub_status: Get http://127.0.0.1:8080/stub_status: dial tcp 127.0.0.1:8080: connect: connection refused

OR

docker run -it -p 9113:9113 nginx/nginx-prometheus-exporter:0.2.0 -nginx.scrape-uri http://127.0.0.1:8080/stub_status
2018/12/05 20:32:15 Starting NGINX Prometheus Exporter Version=0.2.0 GitCommit=ad0a472
2018/12/05 20:32:15 Could not create Nginx Client: Failed to create NginxClient: failed to get http://127.0.0.1:8080/stub_status: Get http://127.0.0.1:8080/stub_status: dial tcp 127.0.0.1:8080: connect: connection refused

It only works if I put the external ip of the server, but how would it be automated?

Thank you in advance.