Desktop Fly

Evgeny Kolyakov — Tue, 15 Sep 2026 15:37:09 +0000

A minimal desktop application for visualizing a fly flying over your desktop and its neuron data using OpenGL. The fly "thinks" in terms of its neuron activity as we stream desktop screenshots and mouse movements into the neural network. The neuron data is taken from male-cns:v1.0 for a male Drosophila (fruit fly) central nervous system 🪰

GitHub: https://github.com/freaker2k7/desktop-fly [MIT]

New Redash fork with AI capabilities!

Evgeny Kolyakov — Thu, 03 Sep 2026 08:03:22 +0000

New Redash features:

Data Source Description: Help the AI to understand your data better.
Text To Query: Use AI to generate SQL, NoSQL and other queries from natural language questions.
Auto Visualizations: Use AI to auto-generate visualizations for questions.
Auto Alerts: Use AI to generate alert suggestions from your data and queries.
Name To Dashboard: Use AI to generate dashboards from your queries and visualizations.
MacOS Support: Now it runs on MacOS as well as on Linux.

DockerServer - Running docker containers from a remote machine in a simple & a secure way.

Evgeny Kolyakov — Sun, 28 Jul 2019 18:51:24 +0000

DockerServer

Super lightweight & simple RESTFul stateless server for running docker containers on a remote machine(s) in a secure way.

Install

npm i -g docker-server

docker run -d -p 1717:1717 --restart=always --name=docker-server -v /var/run/docker.sock:/var/run/docker.sock -e "DS_TOKEN=my_secret_token" evgy/dockerserver

Background

I needed to run a couple of containers on a remote machine and came to these conclusions:

Kubernetes is an overkill !
docker-machine is also complicated !
I just want to run a few containers on a remote machine.

Approach

Built a small REST server with NodeJS, using the express and docker-cli-js packages as a base.

Design Principles

Keep the business logic simple!
It must be stateless!
Docker is (a) present.

Current architecture

The cluster diagram demonstrates a PUT request.

Notes for the Cluster Mode:

*0 - Connection between the load balancer and the docker-server.

*1 - Save the machine load to a JSON file in a shared folder (among all the machines).

PUT method

1 - Requests comes to any free (according to the load balancer) node to answer.

2 - Get the most free (according to actual cpu-mem ratio) node (from the shared storage).

3 - Resend the current request to that node (or process if it's the current node) and return the answer.

For the rest of the methods

Resent the current request to all the nodes and return the merged results.

Usage

Install DockerServer on the machine that you want to run your containers.

DockerServer can be run for a single session with:

$ docker-server

or as a service using PM2:

$ pm2 start /usr/lib/node_modules/docker-server/pm2.config.js

and if you want in addition to start it on startup just run:

$ pm2 startup

And of-course, as mentioned before, but using params, via docker itself:

$ docker run -d -p 1717:1717 --restart=always --name=docker-server -v /var/run/docker.sock:/var/run/docker.sock evgy/dockerserver docker-server --token my_secret_token

Or you can run in HTTPS mode:

(Note that in this example I'm using Let's Encrypt and I'm using readlink because these files are symbolic links)

$ docker run -d -p 443:1717 --privileged --restart=always --name=docker-server -v /var/run/docker.sock:/var/run/docker.sock -v $(readlink -f /home/user/letsencrypt/config/live/your-domain.com/cert.pem):/certs/cert.pem:ro -v $(readlink -f /home/user/letsencrypt/config/live/your-domain.com/chain.pem):/certs/chain.pem:ro -v $(readlink -f /home/user/letsencrypt/config/live/your-domain.com/privkey.pem):/certs/privkey.pem:ro evgy/dockerserver docker-server --token my_secret_token --https

Note: The --privileged argument is only needed in order to use the 443 port, because all ports below 1024 are reserved by root.

Moreover, you can run in a Cluster mode when you have a couple of machines to use:

$ docker run -d -p 1717:1717 --privileged --restart=always --name=docker-server -v /var/run/docker.sock:/var/run/docker.sock -v /some/shared/folder:/my/somewhy/custom/path evgy/dockerserver docker-server --token my_secret_token --cluster --folder /my/somewhy/custom/path

Or simply:

$ docker run -d -p 1717:1717 --privileged --restart=always --name=docker-server -v /var/run/docker.sock:/var/run/docker.sock -v /some/shared/folder:/tmp/docker-server evgy/dockerserver docker-server --token my_secret_token --cluster

Note: /tmp/docker-server is the default folder so you can easily and safely run it even without docker.

Now, you can do "remote" docker operation using simple HTTP requests:

HTTP Method	Endpoint	Desc.	Docker cmd
GET	/	List all the containers	docker ps -a
GET	/:id	Show the logs of a specific container	docker logs :id
PUT	/	Run a container	docker run...
POST	/:id	Execute a command in a container	docker exec...
DELETE	/:id	Delete a container with such a name or an ID	docker rm -f :id

Options

Environment

You can set the following environment variables to configure DockerServer:

Environment Var.	Desc.	Default
`DS_PORT`	The port on which the DockerServer is running	1717
`DS_TOKEN`	The secret token for the authorization	xxxxxx

Parameters

Also, you can start DockerServerwith these parameters:

Param	Desc.	Default
`--port [num]`	The port on which the DockerServer is running	1717
`--token [string]`	The secret token for the authorization	xxxxxx
`--low_burst [num]`	Max number of requests per minute for Low burst.	60
`--mid_burst [num]`	Max number of requests per minute for Mid burst.	180
`--high_burst [num]`	Max number of requests per minute for High burst.	300
`--https`	Enable HTTPS mode. For this you must have the following files: a. /certs/cert.pem b. /certs/privkey.pem c. /certs/chain.pem (optional, to support self-signed certs)	false
`--cluster`	Enable Cluster mode.	false
`--folder [path]`	Shared folder between all docker-servers. (Used only in cluster mode)	/tmp/docker-server
`--cache_interval [num]`	Milliseconds between reads (of all the machines)	3000
`--log_lovel [option]`	Log level [trace\|debug\|info\|warn\|error\|fatal]	info
`--log_expiry [num]`	Time for a log to live in days.	14
`--log_max_size [num]`	Max log size in MB	25
`--help`	Show he
`--version`	Show current version

PUT Data

When sending the PUT request, the following parameters are supported:

Param	Desc.	Default	Docker cmd
image	The image for the run. (required)	null
name	The name of the container.	uuid4()	`--name`
remove	Flag to remove the container when it finishes `--rm`	false	`--rm`
detach	Flag to detach the container `-d`	false	`-d`
ports	Map of ports to publish.	null	`-p`
volumes	Map of volumes to mount.	null	`-v`
data	CMD to run inside the container.	null

POST Data

When sending the POST request, the following parameters are supported:

Param	Desc.	Default	Docker cmd
data	CMD to run inside the container	null
tty	Flag to enable TTY mode	false	`-t`
interactive	Flag to enable interactive mode	false	`-i`

Examples

NOTE: In the examples I assumed you're using the default port.

Get a list of all the containers:

$ curl -X GET http://1.2.3.4:1717/ -H 'Authorization: Basic base64EncodedToken'

Run redis on the remote machine:

$ curl -X PUT http://1.2.3.4:1717/ -H 'Authorization: Basic base64EncodedToken' --data 'name=p-redis&image=redis&ports[1234]=6379'

And/or

$ curl -X PUT http://1.2.3.4:1717/ -H 'Authorization: Basic base64EncodedToken' --data 'name=v-redis&image=redis&volumes[/tmp/data]=/data'

Remove our created container(s):

$ curl -X DELETE http://1.2.3.4:1717/p-redis -H 'Authorization: Basic base64EncodedToken'

And/or

$ curl -X DELETE http://1.2.3.4:1717/v-redis -H 'Authorization: Basic base64EncodedToken'

Changelog

1.8.4 - Edited some docs.

1.8.2 - Added logger.

1.8.1 - Cashed reading (in cluster mode) & fixed host issue in results (also in cluster mode).

1.8.0 - Stable Cluster mode!

See full changelog

Roadmap

Queue (for heavy loads)
Autoscaling

License

APACHE-2.0 (see the LICENSE files in the repository).

Donate

Running dockers is free, but beer is always welcome

or directly donate to our cause

DEV Community: Evgeny Kolyakov