Installation

Requirements

The project leverages docker compose with a custom Bash script and you need to have the following packages installed in your machine:

• docker - v19.03.0+
• docker-compose - v2.3.4+

In some systems you could find pre-installed older versions. Please check this and install a supported version before attempting the installation. Otherwise it would fail. Note: We've added a new Bash script initialize.sh that will check compatibility with your system and attempt to install the required dependencies.

TL;DR

Obviously we strongly suggest reading through all the page to configure ThreatMatrix in the most appropriate way.

However, if you feel lazy, you could just install and test ThreatMatrix with the following steps. docker will be run with sudo if permissions/roles have not been set.

# clone the ThreatMatrix project repository
git clone https://github.com/khulnasoft/ThreatMatrix
cd ThreatMatrix/

# run helper script to verify installed dependencies and configure basic stuff
./initialize.sh

# start the app
./start prod up
# now the application is running on http://localhost:80

# create a super user
sudo docker exec -ti threatmatrix_uwsgi python3 manage.py createsuperuser

# now you can login with the created user from http://localhost:80/login

# Have fun!

Infrastructure Requirements

These are our recommendations for dedicated deployments of ThreatMatrix:

• Basic Installation in a VM: 2 CPU, 4GB RAM, 20GB Disk
• Intensive Usage (hundreds of analysis in a hour) in a single VM: 8CPU, 16GB RAM and 80GB Disk.

Please remember that every environment has its own peculiarities so these numbers must not be taken as the holy grail.

What should be done is a comprehensive evaluation of the environment where the application will deployed.

For more complex environments, a Docker Swarm / Kubernetes cluster is recommended.

ThreatMatrix's maintainers are available to offer paid consultancy and mentorship about that.

Deployment Components

ThreatMatrix is composed of various different technologies, namely:

• React: Frontend, using CRA and certego-ui
• Django: Backend
• PostgreSQL: Database
• Redis: Message Broker
• Celery: Task Queue
• Nginx: Reverse proxy for the Django API and web asssets.
• Uwsgi: Application Server
• Daphne: Asgi Server for WebSockets
• Elastic Search (optional): Auto-sync indexing of analysis' results.
• Kibana (optional): GUI for Elastic Search. We provide a saved configuration with dashboards and visualizations.
• Flower (optional): Celery Management Web Interface

All these components are managed via docker compose.

Deployment Preparation

• Environment configuration (required)
• Database configuration (required)
• Web server configuration (optional)
• Analyzers configuration (optional)

Open a terminal and execute below commands to construct new environment files from provided templates.

./initialize.sh

Environment configuration (required)

In the docker/env_file_app, configure different variables as explained below.

REQUIRED variables to run the image:

• DB_HOST, DB_PORT, DB_USER, DB_PASSWORD: PostgreSQL configuration (The DB credentals should match the ones in the env_file_postgres). If you like, you can configure the connection to an external PostgreSQL instance in the same variables. Then, to avoid to run PostgreSQL locally, please run ThreatMatrix with the option --use-external-database. Otherwise, DB_HOST must be postgres to have the app properly communicate with the PostgreSQL container.

• DJANGO_SECRET: random 50 chars key, must be unique. If you do not provide one, Intel Owl will automatically set a secret key and use the same for each run. The key is generated by initialize.sh script.

  Strongly recommended variable to set:

• THREATMATRIX_WEB_CLIENT_DOMAIN (example: localhost/mywebsite.com): the web domain of your instance, this is used for generating links to analysis results.

Optional configuration:

• OLD_JOBS_RETENTION_DAYS: Database retention for analysis results (default: 14 days). Change this if you want to keep your old analysis longer in the database.

Other optional configuration to enable specific services / features

Configuration required to enable integration with Slack:

• SLACK_TOKEN: Slack token of your Slack application that will be used to send/receive notifications
• DEFAULT_SLACK_CHANNEL: ID of the Slack channel you want to post the message to

Configuration required to have InteOwl sending Emails (registration requests, mail verification, password reset/change, etc)

• DEFAULT_FROM_EMAIL: email address used for automated correspondence from the site manager (example: noreply@mydomain.com)
• DEFAULT_EMAIL: email address used for correspondence with users (example: info@mydomain.com)
• EMAIL_HOST: the host to use for sending email with SMTP
• EMAIL_HOST_USER: username to use for the SMTP server defined in EMAIL_HOST
• EMAIL_HOST_PASSWORD: password to use for the SMTP server defined in EMAIL_HOST. This setting is used in conjunction with EMAIL_HOST_USER when authenticating to the SMTP server.
• EMAIL_PORT: port to use for the SMTP server defined in EMAIL_HOST.
• EMAIL_USE_TLS: whether to use an explicit TLS (secure) connection when talking to the SMTP server, generally used on port 587.
• EMAIL_USE_SSL: whether to use an implicit TLS (secure) connection when talking to the SMTP server, generally used on port 465.

Database configuration (required if running PostgreSQL locally)

If you use a local PostgreSQL instance (this is the default), in the env_file_postgres you have to configure different variables as explained below.

Required variables:

• POSTGRES_PASSWORD (same as DB_PASSWORD)
• POSTGRES_USER (same as DB_USER)
• POSTGRES_DB (default: threat_matrix_db)

Logrotate configuration (strongly recommended)

If you want to have your logs rotated correctly, we suggest you to add the configuration for the system Logrotate. To do that you can leverage the initialize.sh script. Otherwise, if you have skipped that part, you can manually install logrotate by launching the following script:

cd ./docker/scripts
./install_logrotate.sh

We decided to do not leverage Django Rotation Configuration because it caused problematic concurrency issues, leading to logs that are not rotated correctly and to apps that do not log anymore. Logrotate configuration is more stable.

Crontab configuration (recommended for advanced deployments)

We added few Crontab configurations that could be installed in the host machine at system level to solve some possible edge-case issues:

• Memory leaks: Once a week it is suggested to do a full restart of the application to clean-up the memory used by the application. Practical experience suggest us to do that to solve some recurrent memory issues in Celery. A cron called application_restart has been added for this purpose (it uses the absolute path of start script in the container). This cron assumes that you have executed ThreatMatrix with the parameters --all_analyzers. If you didn't, feel free to change the cron as you wish.

This configuration is optional but strongly recommended for people who want to have a production grade deployment. To install it you need to run the following script in each deployed server:

cd ./docker/scripts
./install_crontab.sh

Web server configuration (required for enabling HTTPS)

Intel Owl provides basic configuration for:

• Nginx (configuration/nginx/http.conf)
• Uwsgi (configuration/threat_matrix.ini)

In case you enable HTTPS, remember to set the environment variable HTTPS_ENABLED as "enabled" to increment the security of the application.

There are 3 options to execute the web server:

• HTTP only (default)

  The project would use the default deployment configuration and HTTP only.

• HTTPS with your own certificate

  The project provides a template file to configure Nginx to serve HTTPS: configuration/nginx/https.conf.

  You should change ssl_certificate, ssl_certificate_key and server_name in that file and put those required files in the specified locations.

  Then you should call the ./start script with the parameter --https to leverage the right Docker Compose file for HTTPS.

  Plus, if you use Flower, you should change in the docker/flower.override.yml the flower_http.conf with flower_https.conf.

• HTTPS with Let's Encrypt

  We provide a specific docker-compose file that leverages Traefik to allow fast deployments of public-faced and HTTPS-enabled applications.

  Before using it, you should configure the configuration file docker/traefik.override.yml by changing the email address and the hostname where the application is served. For a detailed explanation follow the official documentation: Traefix doc.

  After the configuration is done, you can add the option --traefik while executing ./start

Run

Now that you have completed different configurations, starting the containers is as simple as invoking:

$ ./start prod up

You can add the docker options -d to run the application in the background.

./start prod up -- -d

docker compose --project-directory docker -f docker/default.yml -f docker/postgres.override.yml -f docker/redis.override.yml -f docker/nginx.override.yml -p threatmatrix up

Stop

To stop the application you have to:

• if executed without -d parameter: press CTRL+C
• if executed with -d parameter: ./start prod down

Cleanup of database and application

This is a destructive operation but can be useful to start again the project from scratch

./start prod down -- -v

After Deployment

Users creation

You may want to run docker exec -ti threatmatrix_uwsgi python3 manage.py createsuperuser after first run to create a superuser. Then you can add other users directly from the Django Admin Interface after having logged with the superuser account. To manage users, organizations and their visibility please refer to this section

Update and Rebuild

Rebuilding the project / Creating custom docker build

If you make some code changes and you like to rebuild the project, follow these steps:

1. ./start test build -- --tag=<your_tag> . to build the new docker image.
2. Add this new image tag in the docker/test.override.yml file.
3. Start the containers with ./start test up -- --build.

Update to the most recent version

To update the project with the most recent available code you have to follow these steps:

$ cd <your_threat_matrix_directory> # go into the project directory
$ git pull # pull new changes
$ ./start prod down # kill and destroy the currently running ThreatMatrix containers
$ ./start prod up # restart the ThreatMatrix application

PermissionError: [Errno 13] Permission denied: '/var/log/threat_matrix/django/authentication.log

sudo chown -R www-data:www-data /var/lib/docker/volumes/threat_matrix_generic_logs/_data/django

Updating to >=6.0.0 from a 5.x.x version

ThreatMatrix v6 introduced some major changes regarding how the project is started. Before upgrading, some important things should be checked by the administrator:

• Docker Compose V1 support has been dropped project-wide. If you are still using a Compose version prior to v2.3.4, please upgrade to a newer version or install Docker Compose V2.
• ThreatMatrix is now started with the new Bash start script that has the same options as the old Python start.py script but is more manageable and has decreased the overall project dependencies. The start.py script has now been removed. Please use the new start script instead.
• The default message broker is now Redis. We have replaced Rabbit-MQ for Redis to allow support for Websockets in the application:
  • This change is transparent if you use our start script to run ThreatMatrix. That would spawn a Redis instance instead of a Rabbit-MQ one locally.
  • If you were using an external broker like AWS SQS or a managed Rabbit-MQ, they are still supported but we suggest to move to a Redis supported service to simplify the architecture (because Redis is now mandatory for Websockets)
• Support for multiple jobs with multiple playbooks has been removed. Every Observable or File in the request will be processed by a single playbook.
• We upgraded the base PostgreSQL image from version 12 to version 16. You have 2 choice:
  • remove your actual database and start from scratch with a new one
  • maintain your database and do not update Postgres. This could break the application at anytime because we do not support it anymore.
  • if you want to keep your old DB, follow the migration procedure you can find below

The database migration procedure is as follows:

• You have ThreatMatrix version 5.x.x up and running
• Bring down the application (you can use the start script or manually concatenate your docker compose configuration )
• Go inside the docker folder cd docker
• Bring only the postgres 12 container up docker run -d --name threatmatrix_postgres_12 -v threat_matrix_postgres_data:/var/lib/postgresql/data/ --env-file env_file_postgres library/postgres:12-alpine
• Dump the entire database. You need the user and the database that you configured during startup for this docker exec -t threatmatrix_postgres_12 pg_dump -U <POSTGRES_USER> -d <POSTGRES_DB> --no-owner > /tmp/dump_threatmatrix.sql
• Stop che container docker container stop threatmatrix_postgres_12
• Remove the backup container docker container rm threatmatrix_postgres_12
• Remove the postgres volume docker volume rm threat_matrix_postgres_data <------------- remove old data, this is not exactly necessary because the new postgres has a different volume name
• Start the intermediary postgres 16 container docker run -d --name threatmatrix_postgres_16 -v threatmatrix_postgres_data:/var/lib/postgresql/data/ --env-file env_file_postgres library/postgres:16-alpine
• Add the data to the volume cat /tmp/dump_threatmatrix.sql | docker exec -i threatmatrix_postgres_16 psql -U <POSTGRES_USER> -d <POSTGRES_DB>
• Stop the intermediary container docker container stop threatmatrix_postgres_16
• Remove the intermediary container docker container rm threatmatrix_postgres_16
• Update ThreatMatrix to the latest version
• Bring up the application back again (you can use the start script or manually concatenate your docker compose configuration)

Updating to >=5.0.0 from a 4.x.x version

ThreatMatrix v5 introduced some major changes regarding how the plugins and their related configuration are managed in the application. Before upgrading, some important things should be checked by the administrator:

• A lot of database migrations will need to be applied. Just be patient few minutes once you install the new major release. If you get 500 status code errors in the GUI, just wait few minutes and then refresh the page.
• We moved away from the old big analyzer_config.json which was storing all the base configuration of the Analyzers to a database model (we did the same for all the other plugins types too). This allows us to manage plugins creation/modification/deletion in a more reliable manner and via the Django Admin Interface. If you have created custom plugins and changed those <plugins>_config.json file manually, you would need to re-create those custom plugins again from the Django Admin Interface. To do that please follow the related new documentation
• We have REMOVED all the analyzers that we deprecated during the v4 releases cycle. Please substitute them with their respective new names, in case they have a replacement.
  • REMOVED Pulsedive_Active_IOC analyzer. Please substitute it with the new Pulsedive analyzer.
  • REMOVED Fortiguard analyzer because endpoint does not work anymore. No substitute.
  • REMOVED Rendertron analyzer not working as intended. No substitute.
  • REMOVED ThreatMiner, SecurityTrails and Robtex various analyzers and substituted with new versions.
  • REMOVED Doc_Info_Experimental. Its functionality (XLM Macro parsing) is moved to Doc_Info
  • REMOVED Strings_Info_Classic. Please use Strings_Info
  • REMOVED Strings_Info_ML. Please use Strings_Info and set the parameter rank_strings to True
  • REMOVED all Yara_Scan_<repo> analyzers. They all went merged in the single Yara analyzer
  • REMOVED Darksearch_Query analyzer because the service does not exist anymore. No substitute.
  • REMOVED UnpacMe_EXE_Unpacker. Please use UnpacMe
  • REMOVED BoxJS_Scan_JavaScript. Please use BoxJS
  • REMOVED all Anomali_Threatstream_<option> analyzers. Now we have a single Anomali_Threatstream analyzer. Use the parameters to select the specific API you need.

Updating to >=5.0.0 from a 3.x.x version

This is not supported. Please perform a major upgrade once at a time.

Updating to >=4.0.0 from a 3.x.x version

ThreatMatrix v4 introduced some major changes regarding the permission management, allowing an easier way to manage users and visibility. But that did break the previous available DB. So, to migrate to the new major version you would need to delete your DB. To do that, you would need to delete your volumes and start the application from scratch.

python3 start.py prod down -v

Please be aware that, while this can be an important effort to manage, the v4 ThreatMatrix provides an easier way to add, invite and manage users from the application itself. See the Organization section.

Updating to >=2.0.0 from a 1.x.x version

Users upgrading from previous versions need to manually move env_file_app, env_file_postgres and env_file_integrations files under the new docker directory.

Updating to >v1.3.x from any prior version

If you are updating to >v1.3.0 from any prior version, you need to execute a helper script so that the old data present in the database doesn't break.

1. Follow the above updation steps, once the docker containers are up and running execute the following in a new terminal

   docker exec -ti threatmatrix_uwsgi bash
   

   to get a shell session inside the ThreatMatrix's container.

2. Now just copy and paste the below command into this new session,

   curl https://gist.githubusercontent.com/Eshaan7/b111f887fa8b860c762aa38d99ec5482/raw/758517acf87f9b45bd22f06aee57251b1f3b1bbf/update_to_v1.3.0.py | python -
   

3. If you see "Update successful!", everything went fine and now you can enjoy the new features!