Before starting the installation, let's discover how OpenCTI is working, which dependencies are needed and what are the minimal requirements to deploy it in production.
The OpenCTI platform relies on several external databases and services in order to work.
The platform is the central part of the OpenCTI technological stack. It allows users to access to the user interface but also provides the GraphQL API used by connectors and workers to insert data. In the context of a production deployment, you may need to scale horizontally and launch multiple platforms behind a load balancer connected to the same databases (ElasticSearch, Redis, S3, RabbitMQ).
The workers are standalone Python processes consuming messages from the RabbitMQ broker in order to do asynchronous write queries. You can launch as many workers as you need to increase the write performances. At some point, the write performances will be limited by the throughput of the ElasticSearch database cluster.
Number of workers
If you need to increase performances, it is better to launch more platforms to handle worker queries. The recommended setup is to have at least one platform for 3 workers (ie. 9 workers distributed over 3 platforms).
The connectors are third-party pieces of software (Python processes) that can play five different roles on the platform:
|Pull data from remote sources, convert it to STIX2 and insert it on the OpenCTI platform.
|MITRE Datasets, MISP, CVE, AlienVault, Mandiant, etc.
|Listen for new OpenCTI entities or users requests, pull data from remote sources to enrich.
|Shodan, DomainTools, IpInfo, etc.
|Extract data from files uploaded on OpenCTI trough the UI or the API.
|STIX 2.1, PDF, Text, HTML, etc.
|Generate export from OpenCTI data, based on a single object or a list.
|STIX 2.1, CSV, PDF, etc.
|Consume a platform data stream an do something with events.
|Splunk, Elastic Security, Q-Radar, etc.
List of connectors
You can find all currently available connector in the OpenCTI Ecosystem.
|S3 / MinIO
To have more details about deploying OpenCTI and its dependencies in cluster mode, please read the dedicated section.