Running a local kibana in a container - docker

I am trying to run use kibana console with my local elasticsearch (container)
In the ElasticSearch documentation I see
docker run -p 9200:9200 -p 9300:9300 -e "discovery.type=single-node" docker.elastic.co/elasticsearch/elasticsearch:6.2.2
Which lets me run the community edition in a quick one liner.
Looking at the kibana documentation i see only
docker pull docker.elastic.co/kibana/kibana:6.2.2
Replacing pull with run it looks for the x-pack (I think it means not community) and fails to find the ES
Unable to revive connection: http://elasticsearch:9200/
Is there a one liner that could easily set up kibana localy in a container?
All I need is to work with the console (Sense replacement)

If you want to use kibana with elasticsearch locally with docker, they have to communicate with each other. To do so, according to the doc, you need to link the containers.
You can give a name to the elasticsearch container with --name:
docker run \
--name elasticsearch_container \
--publish 9200:9200 \
--publish 9300:9300 \
--env "discovery.type=single-node" \
docker.elastic.co/elasticsearch/elasticsearch:6.2.2
And then link this container to kibana:
docker run \
--name kibana \
--publish 5601:5601 \
--link elasticsearch_container:elasticsearch_alias \
--env "ELASTICSEARCH_URL=http://elasticsearch_alias:9200" \
docker.elastic.co/kibana/kibana:6.2.2
The port 5601 is exposed locally to access it from your browser. You can check in the monitoring section that elasticsearch's health is green.
EDIT (24/03/2020):
The option --link may eventually be removed and is now a legacy feature of docker.
The idiomatic way of reproduce the same thing is to firstly create a user-defined bridge:
docker network create elasticsearch-kibana
And then create the containers inside it:
 Version 6
docker run \
--name elasticsearch_container \
--network elasticsearch-kibana \
--publish 9200:9200 \
--publish 9300:9300 \
--env "discovery.type=single-node" \
docker.elastic.co/elasticsearch/elasticsearch:6.2.2
docker run \
--name kibana \
--publish 5601:5601 \
--network elasticsearch-kibana \
--env "ELASTICSEARCH_URL=http://elasticsearch_container:9200" \
docker.elastic.co/kibana/kibana:6.2.2
Version 7
As it was pointed out, the environment variable changed for the version 7. It now is ELASTICSEARCH_HOSTS.
docker run \
--name elasticsearch_container \
--network elasticsearch-kibana \
--publish 9200:9200 \
--publish 9300:9300 \
--env "discovery.type=single-node" \
docker.elastic.co/elasticsearch/elasticsearch:7.6.2
docker run \
--name kibana \
--publish 5601:5601 \
--network elasticsearch-kibana \
--env "ELASTICSEARCH_HOSTS=http://elasticsearch_container:9200" \
docker.elastic.co/kibana/kibana:7.6.2
User-defined bridges provide automatic DNS resolution between containers that means you can access each other by their container names.

It is convenient to use docker-compose as well.
For instance, the file below, stored in home directory, allows to start Kibana with one command:
docker-compose up -d:
# docker-compose.yml
version: "2"
kibana:
image: "docker.elastic.co/kibana/kibana:6.2.2"
container_name: "kibana"
environment:
- "ELASTICSEARCH_URL=http://<elasticsearch-endpoint>:9200"
- "XPACK_GRAPH_ENABLED=false"
- "XPACK_ML_ENABLED=false"
- "XPACK_REPORTING_ENABLED=false"
- "XPACK_SECURITY_ENABLED=false"
- "XPACK_WATCHER_ENABLED=false"
ports:
- "5601:5601"
restart: "unless-stopped"
In addition, Kibana service might be a part of your project in development environment (in case, docker-compose is used).

Related

How to access Gitlab's metrics (Prometheus and Grafana) from Docker installation?

I installed Gitlab using Docker image on a Ubuntu virtual machine running on a MAC M1 as follows (https://hub.docker.com/r/yrzr/gitlab-ce-arm64v8):
docker run \
--detach \
--restart always \
--name gitlab-ce \
--privileged \
--memory 4096M \
--publish 22:22 \
--publish 80:80 \
--publish 443:443 \
--hostname 127.0.0.1 \
--env GITLAB_OMNIBUS_CONFIG=" \
nginx['redirect_http_to_https'] = true; "\
--volume /srv/gitlab-ce/conf:/etc/gitlab:z \
--volume /srv/gitlab-ce/logs:/var/log/gitlab:z \
--volume /srv/gitlab-ce/data:/var/opt/gitlab:z \
yrzr/gitlab-ce-arm64v8:latest
All seems to be working correctly on localhost, except that I can't access the metrics, I got unable to connect error on:
Prometheus: http://localhost:9090
Grafana: http://localhost/-/grafana
I tried enabling metrics as in the documentation, and docker exec -it gitlab-ce gitlab-ctl reconfigure
What I'm missing?
Thanks
When Gitlab uses localhost this will resolve the localhost on the container and not the host (so your Mac).
There are two options to solve this:
Use host.docker.internal instead of localhost (this resolves to the internal IP address used by the host) - see this doc for more info
Configure your container to use the host network by adding this to the docker run command: --network=host which will let your container and host to share the same network stack (however, this is not supported nu Docker Desktop for mac according to this)

How to connect my Kibana to ElasticSearch in the docker run command?

Just trying to learn to setup Kibana and Elastic search using native docker command (i.e. not using Docker-Compose).
Below are the commands I run
docker network create es-net
docker run -d --name es-cluster \
--net es-net -p 9200:9200 \
-e "xpack.security.enabled=false" \
-e "discovery.type=single-node" \
docker.elastic.co/elasticsearch/elasticsearch:7.2.0
docker run -d --net es-net -p 5601:5601 \
-e ELASTICSEARCH_URL=http://es-cluster:9200 \
docker.elastic.co/kibana/kibana:7.2.0
Somehow Kibana is not loading the elastic search up when I run http://localhost:5601/ and always with the message Kibana server is not ready yet
I follow the answer as per Kibana on Docker cannot connect to Elasticsearch, to ensure the ELASTICSEARCH_URL is correctly set, but it is still not coming up. Anything I miss?
note: tested with curl 0.0.0.0:9200, the elastic search is already running
Looks like since I'm in version 7.2.0 of Kibana, it has changed from ELASTICSEARCH_URL to ELASTICSEARCH_HOSTS
as per https://www.elastic.co/guide/en/kibana/current/docker.html
docker run -d --net es-net -p 5601:5601 \
-e ELASTICSEARCH_HOSTS=http://es-cluster:9200 \
docker.elastic.co/kibana/kibana:7.2.0
With this in place, all should work then.

Binding (sub)domain to traefik dashboard has 503 error

Running traefik with docker and bind (sub)domain to dashboard but i got 503 when request to that.
Traefik is modern proxy reverse and i run it with docker by blew code. to see dashboard without direct url i binding port 8080 to 8080 and after that i can see dashboard. in dashboard Route Rule Host:monitor.monitor.my_domain was present in front and http://172.20.0.3:8080 was present in back but when i try to access to http://monitor.my_domain i got 503 error
docker run -d \
-v /var/run/docker.sock:/var/run/docker.sock \
-v $PWD/traefik.toml:/traefik.toml \
-v $PWD/acme.json:/acme.json \
-p 80:80 \
-p 443:443 \
-l traefik.frontend.rule=Host:monitor.my_domain\
-l traefik.port=8080 \
--network web \
--name traefik \
traefik:1.7.6-alpine
You did not post your traefik.toml and overall your question is very difficult to parse. To me it sounds unlikely that without more information anyone is able to help.
If you make sure that monitor.my_domain resolves to your traefik instance and then run:
docker run -d \
-v /var/run/docker.sock:/var/run/docker.sock \
-p 80:80 \
-p 443:443 \
-l traefik.frontend.rule=Host:monitor.my_domain\
-l traefik.port=8080 \
--name traefik \
traefik:1.7.6-alpine --api --docker
This will work. Navigate in your browser to http://monitor.my_domain (that you made sure to resolve to traefik) and you will see the dashboard.
My advice is to try that first, make sure that it works, and if it does, try to find out what's different with your own setup, by tweaking slightly this example or your own until either this one breaks or yours work.

Build from linuxserver\deluge

I'd like to be able to use a Dockerfile with the linuxserver\deluge image but I'm unsure what is the correct way to do this in a docker-compose.yaml file.
docker create \
--name=deluge \
--net=host \
-e PUID=1001 \
-e PGID=1001 \
-e UMASK_SET=<022> \
-e TZ=<timezone> \
-v </path/to/deluge/config>:/config \
-v </path/to/your/downloads>:/downloads \
--restart unless-stopped \
linuxserver/deluge
Can someone help me convert this please so that I can use a Dockerfile
Thanks :)
The following docker-compose.yml file is similar to your command :
version: "3"
services:
deluge:
container_name: deluge
image: linuxserver/deluge
environment:
- PUID=1001
- PGID=1001
- UMASK_SET=<022>
- TZ=<timezone>
volumes:
- </path/to/deluge/config>:/config
- </path/to/your/downloads>:/downloads
restart: unless-stopped
network_mode: host
Documentation is a great place to find the mapping between docker options and docker-compose syntax. Here is a recap of what have been used for this example :
--name => container_name
-e => environment (array of key=value)
-v => volumes (array of volume_or_folder_on_host:/path/inside/container)
--restart <policy> => restart: <policy>
--net=xxxx => network_mode
You can now run docker-compose up to start all your services (only deluge here) instead of your docker run command.

Docker Swarm Windows Worker with Traefik returns Gateway Timeout

The objective is to get a mixed OS Docker swarm running using Linux servers and Windows 10 Machines running Docker For Windows
Currently Windows workers are theoretically supported on mixed os swarms provided the --endpoint-mode flag is set to 'dnsrr'. This is explained here. However attempts to use traefik to route to a simple docker whoami image stefanscherer/whoami image have failed.
Minimal Failing Example
// On (Linux) Manager Node:
docker swarm init --advertise-addr <hostaddress> --listen-addr <hostaddress>:2377
// On (Windows 10) Worker Node:
docker swarm join <jointoken>
// On Manager Node:
docker network create --driver=overlay traefik-net
docker service create \
--name traefik \
--constraint=node.role==manager \
--publish 80:80 --publish 8080:8080 \
--mount type=bind,source=/var/run/docker.sock,target=/var/run/docker.sock \
--network traefik-net \
traefik \
--docker \
--docker.swarmmode \
--docker.domain=traefik \
--docker.watch \
--web
docker service create \
--name whoami \
--label traefik.enable=true \
--label traefik.frontend.rule=Host:whoami.docker \
--label traefik.protocol=http \
--label traefik.docker.network=traefik-net \
--label traefik.backend.loadbalancer.method=drr \
--label traefik.backend=whoami \
--network traefik-net \
--mode global\
--label traefik.port=80 \
stefanscherer/whoami
Traefik successfully sets up backend rules, to check the routing I used the traefik dashboard to find out the URL that is routed to by the rule e.g. '10.0.0.12:8080'. I then compare this with the IP address of each task, the task can be viewed with docker service ps, and their address' found using
docker inspect <taskID> \
--format '{{ range .NetworksAttachments }}{{ .Addresses }}{{ end }}'
The Problem
A HTTP request with a header 'Host:whoami.docker' sent to the IP of the manager will succeed when routed to the manager and fail with 504 Gateway Timeout when routed to the Windows Task on the Windows worker.
You're missing setting --endpoint-mode=dnsrr to your whoami service.
docker service create \
--name whoami \
--label traefik.enable=true \
--label traefik.frontend.rule=Host:whoami.docker \
--label traefik.protocol=http \
--label traefik.docker.network=traefik-net \
--label traefik.backend.loadbalancer.method=drr \
--label traefik.backend=whoami \
--network traefik-net \
--mode global\
--label traefik.port=80 \
--endpoint-mode=dnsrr
stefanscherer/whoami
Setting endpoint-mode dnsrr will disable VIP address which probably is causing the issue.
I had the same problem when using the stefanscherer/whoami image. Using microsoft/dotnet-samples:aspnetapp works though, so the error seems related to the image.
I'm using the following setup:
Ubuntu 16.04
Docker 18.03.1-ce
Run as Manager
Runs traefik
Windows 1803
Docker 18.03.1-ee-2
Runs as Worker (joining as Manager did not work)
Runs microsoft/dotnet-samples:aspnetapp

Resources