# Review deployment requirements

Supported in: This document outlines the technical and environmental requirements necessary
to deploy a Google Security Operations Remote Agent. The Remote Agent acts as a bridge,
letting your Google Security Operations platform communicate with security tools and
devices within your local environment.

## Before you begin

Before proceeding with the installation, confirm that you meet the following prerequisites:

- **Email configuration (optional)** : Configure the email (SMTP/IMAP) integration in the Content Hub only if you require the installation package to be sent by email. If configured, use **Shared Instances** on the **Configure** page.
- **Server accessibility**: The Agent server must be accessible to both the Google SecOps Publisher and the target security tools within the corresponding environment.

### Technical requirements

Review the following hardware and operating system requirements for the server hosting the Remote Agent:

| Requirement | Basic deployment | Scaled-up deployment |
|---|---|---|
| CPU | 4 cores | 8 cores |
| RAM | 8 GB | 16 GB |
| Storage | 100 GB | 100 GB |

## Supported operating systems

Following are the prerequisites for the end customer server (for Remote Agent manual
installation only):

- Clean Centos 7.9
- Red Hat Enterprise Linux (RHEL) 8.7
- Debian 12

If your infrastructure requires RHEL 9 or 10 (or
later), deploy the Remote Agent using a supported container engine
([Podman](https://docs.cloud.google.com/chronicle/docs/soar/working-with-remote-agents/deploy-agent-with-podman)
or [Docker](https://docs.cloud.google.com/chronicle/docs/soar/working-with-remote-agents/deploy-agent-with-docker))
rather than the manual installer.

## Supported container engines

The Remote Agent supports the following container engines:

- **Podman** (recommended for all new installations)
- **Docker**

Containerized deployments run a prebuilt container image (based on Debian 12
for Remote Agent versions later than 2.5.0.4) and are supported on hosts running
RHEL 9 or 10 (or later), as well as RHEL 8 and Debian 12.

## Network and firewall requirements

When you configure firewall rules for the Remote Agent host, meet the
following outbound connectivity requirements:

- Allow outbound HTTPS traffic on TCP port `443` to your Google SecOps instance, Publisher, and locational API endpoints. For the list of required domains, see [Network access and domain allowlist](https://docs.cloud.google.com/chronicle/docs/log-in-to-ui#network-access-and-domain-allowlist). If your firewall policy doesn't allow wildcard domain entries (such as `*.backstory.chronicle.security`, `*.siemplify-soar.com`, or `*-chronicle.googleapis.com`), use the tenant-specific SOAR URL, Publisher hostname (`SERVER_API_ROOT`), and locational API endpoint (`<var>REGION</var>-chronicle.googleapis.com`, configured in `ONE_PLATFORM_URL_DOMAIN`) provided when your tenant is provisioned. The regional endpoint format (`chronicle.<var>REGION</var>.rep.googleapis.com`) isn't supported for Remote Agent control-plane connectivity.
- Allow outbound HTTPS traffic on TCP port `443` to the required Google Cloud authentication endpoints: `oauth2.googleapis.com` (for service account authentication), `sts.googleapis.com`, and `iamcredentials.googleapis.com` (for [Workload Identity Federation](https://docs.cloud.google.com/chronicle/docs/soar/working-with-remote-agents/authenticate-remote-agent-with-workload-identity-federation)). For more information, including Private Service Connect configuration, see [Network and Private Service Connect requirements](https://docs.cloud.google.com/chronicle/docs/soar/working-with-remote-agents/migrate-remote-agent-to-google#network-and-psc-requirements).
- Allow outbound HTTPS traffic on TCP port `443` to the container registries used for Podman or Docker deployments. Required registries include `*.pkg.dev` (such as `us-docker.pkg.dev`) and `gcr.io` for Artifact Registry, or `docker.io`.
- Bypass SSL/TLS inspection on application-layer proxies and Deep Packet Inspection (DPI) firewalls for Remote Agent traffic. This prevents proxies from interrupting TLS handshakes with your Publisher and Google SecOps API endpoints.

## Hardened environments and OS security policies

Host OS hardening and security enforcement mechanisms---such as **SELinux** in
`Enforcing` mode, **AppArmor**, or endpoint security software---can cause Remote
Agent tasks to fail when the security policy blocks OS commands executed by the
agent.

If you deploy the Remote Agent on a hardened host, you must exempt the Remote
Agent from these security enforcement policies so that it can run required OS
commands:

- For containerized deployments (Docker or Podman), exempt the Remote Agent container when starting it for the first time by adding `--security-opt label=disable` (for SELinux) or `--security-opt apparmor=unconfined` (for AppArmor) to your `docker run` or `podman run` command.
- Configure custom SELinux policy rules, AppArmor profiles, or endpoint security and antivirus exemptions for the Remote Agent directories (`/opt/SiemplifyAgent`, `/home/siemplify_agent/agent_source`, and `/var/log/SiemplifyAgent`) and executables (`/usr/local/bin/python3.11`, `/usr/local/bin/pip3`, and `/usr/bin/supervisord`).
- Alternatively, set SELinux to `Permissive` mode on the host machine by setting `SELINUX=permissive` in `/etc/selinux/config` (or temporarily running `sudo setenforce 0`) if targeted policy exemptions aren't configured.

**Need more help?** [Get answers from Community members and Google SecOps professionals.](https://security.googlecloudcommunity.com/)