Cisco Threat Grid Appliance Administrator Guide Version 2.12
Bias-Free Language
The documentation set for this product strives to use bias-free language. For the purposes of this documentation set, bias-free is defined as language that does not imply discrimination based on age, disability, gender, racial identity, ethnic identity, sexual orientation, socioeconomic status, and intersectionality. Exceptions may be present in the documentation due to language that is hardcoded in the user interfaces of the product software, language used based on RFP documentation, or language that is used by a referenced third-party product. Learn more about how Cisco is using Inclusive Language.
The Cisco Threat Grid Appliance is a Linux server with Threat Grid software installed by
Cisco Manufacturing prior to shipment. Once a new Threat Grid Appliance is received, it
must be set up and configured for your on-premises network environment.
This chapter describes the environmental, hardware, and network requirements that should
be reviewed prior to setup and configuration:
Supported Browsers
Threat Grid supports the following browsers:
Google Chrome™
Mozilla Firefox®
Apple Safari®
Note
Microsoft Internet Explorer is not supported.
Environmental Requirements
Threat Grid Appliance (v2.7.2 and later) is deployed on the Threat Grid M5 Appliance
server. Before you set up and configure the Threat Grid Appliance, make sure the
necessary environmental requirements for power, rack space, cooling, and other issues
are met, according to the specifications in the Cisco Threat Grid M5 Hardware Installation
Guide.
Hardware Requirements
The SFP+ form factor is used for the Admin interface. If you are clustering Threat Grid
Appliances, each one will require an additional SFP+ module on the Clust interface.
Note
The SFP+ modules must be connected before the Threat Grid Appliance is powered
on for the session in which the configuration wizard is going to be run.
If there are no SFP+ ports available on the switch, or SFP+ is not desirable, then a
transceiver for 1000Base-T can be used (for example, Cisco Compatible Gigabit RJ 45
Copper SFP Transceiver Module Mini -GBIC - 10/100/1000 Base-T Copper SFP Module).
You can attach a monitor to the server, or, if Cisco Integrated Management Controller
(CIMC) is configured, you can use a remote KVM (on UCS C220-M3 and C220-M4 servers).
Note
CIMC is not supported on the Threat Grid M5 Appliance server.
The Threat Grid Appliance requires three networks:
ADMIN - The Administrative network must be configured to perform the
Threat Grid Appliance setup.
Admin UI Management Traffic (HTTPS)
SSH
NFSv4 (Outbound. If a NFS hostname is used instead of IP, this name will
be resolved via Dirty DNS.)
CLEAN - The Clean network is used for inbound, trusted traffic to the
Threat Grid Appliance (requests), and integrated appliances such as the Cisco
Email Security Appliance and Web Security Appliance; integrated appliances
connect to the IP address of the Clean interface.
Note
The URL for the Clean network interface will not work until the Admin UI
configuration is complete.
The following specific, restricted types of network traffic can be outbound from
the Clean network:
Remote syslog connections
Email messages sent by the Threat Grid Appliance
Disposition Update Service connections to AMP for Endpoints Private Cloud
devices
DNS requests (related to any of the above)
LDAP
DIRTY - The Dirty network is used for outbound traffic from the Threat
Grid Appliance (including malware traffic).
Note
To protect your internal network assets, we recommend using a dedicated
external IP address (for example, the Dirty interface) that is different
from your corporate IP.
The DNS server needs to be accessible via the Dirty network when used for purposes other
than Disposition Update Service lookups, resolving remote syslog connections, and
resolving the mail server used for notifications from the Threat Grid software.
By default, DNS uses the Dirty interface. The Clean interface is used for AMP for
Endpoints Private Cloud integrations. If the AMP for Endpoints Private Cloud hostname
cannot be resolved over the Dirty interface, then a separate DNS server that uses the
Clean interface can be configured in the Admin UI.
NTP Server Access
The NTP server needs to be accessible via the Dirty network.
Starting with the 2.12 release, an appliance can be optionally configured to connect to
an NTP server from the clean interface rather than the dirty interface (default). This
makes it possible to use an internal NTP server.
Integrations
Additional planning may be required if the Threat Grid Appliance is going to be used with
other Cisco products, such as the Email Security Appliance, Web Security Appliance, or
AMP for Endpoints Private Cloud. See Connecting ESA/WSA to Threat Grid
Appliance for more information.
DHCP Requirements
If you are connected to a network configured to use DHCP, it is important that you
understand the requirements. Threat Grid Appliances that use DHCP need to explicitly
specify DNS.
Warning
An upgrade of a system without a DNS server explicitly specified will fail.
Note
The TGSH Dialog displays the information you will need to access and configure the
Admin UI. It may take some time for the IP addresses for DHCP to display after your
appliance boots.
Open the TGSH Dialog and note the following information:
Admin URL - The Admin network. You will need this address in order to
continue the remaining configuration tasks in the Admin UI.
Application URL - The Clean network. This is the address to use after
completing the configuration in the Admin UI.
The Dirty network is not shown.
Password - The initial Admin password that is randomly generated during
the Threat Grid Appliance installation. You will need to change this password
later as the first step the Admin UI configuration process.
If you need to change your initial IP assignments from DHCP to static IP addresses, see
Configuring Network and DHCP.
License
When a new appliance is purchased, a license is generated and the Retrieve License
From Server button on the Configuration > License page is enabled.
However, if that doesn't work or if there's a special case (such as a license being a
custom one-off), then you will be given the license directly, as an encrypted file with
a password.
For additional questions about licenses, contact Support.
Rate Limits
The API sample submission rate limit is global for the Threat Grid Appliance under the
terms of the license agreement. This affects API submissions ONLY, not manual sample
submissions.
Rate limits are based on a window of rolling time, not to a calendar day. When the
submission limit is exhausted, the next API submission will return a 429 error, plus a
message about how long to wait before retrying. See the portal online Help for more
information.
Organizations and Users
Once you have completed the Threat Grid Appliance setup and network configuration, you
must create the initial Threat Grid organizations and add user account(s), so that
people can login and begin submitting malware samples for analysis. This task may
require planning and coordination among multiple organizations and users, depending on
your requirements.
See Create New Organizations and the Threat Grid portal
Help (click Administration > Administrator's Guide to open the Administration
Guide topic) for additional information.
Updates
The initial Threat Grid Appliance setup and configuration steps must be completed
before installing any Threat Grid Appliance updates. We recommend that you check for
updates immediately after completing the initial configuration (see the Cisco Threat Grid Appliance Getting Started
Guide).
Threat Grid Appliance updates cannot be downloaded until the license is installed, and
the update process requires that the initial appliance configuration is completed.
Updates must be done in sequence.
User Interfaces
After the server has been correctly attached to the network and powered up, there are
several user interfaces available for configuring the Threat Grid Appliance.
Note
LDAP authentication is available for TGSH Dialog and the Admin UI. RADIUS
authentication is available for the Threat Grid Application UI (v2.10 and
later).
TGSH Dialog
The TGSH Dialog interface is used to configure the network interfaces. The TGSH
Dialog is displayed when the Threat Grid Appliance successfully boots up.
Reconnecting to the TGSH Dialog
The TGSH Dialog remains open on the console and can be accessed either by attaching a
monitor to the appliance or, if CIMC is configured, via remote KVM.
Note
CIMC is not supported on the Threat Grid M5 Appliance server.
To reconnect to the TGSH Dialog, ssh into the Admin IP address as the user
threatgrid.
The required password is either the initial, randomly generated password, which is
visible initially in the TGSH Dialog, or the new Admin password you create during
the first step of the Admin UI Configuration (see the Cisco Threat Grid Appliance Getting Started
Guide).
Threat Grid Shell (tgsh)
The Threat Grid Shell (tgsh) is an administrator's interface that is used to execute
commands (including destroy-data and forced backup), and for expert, low-level
debugging. To access tgsh, choose CONSOLE in the TGSH Dialog.
Note
The Admin UI uses the same credentials as the Threat Grid user, so any password
changes/updates made via tgsh will also impact the Admin UI.
Caution
Network configuration changes made with tgsh are not supported unless specifically
directed by Threat Grid support; the Admin UI or TGSH Dialog should be used instead.
Options to modify admin email, glovebox URL, SMTP configuration, etc. have been
removed with the 2.12 release. The Wipe Appliance operation is now activated within
recovery mode tgsh rather than the bootloader menu.
Admin UI
This is the primary Threat Grid user interface used for configuration. Much of the Threat
Grid Appliance configuration can ONLY be done via the Admin UI, including licenses,
email host, and SSL certificates.
Threat Grid Portal
The Threat Grid user interface application is available as a cloud service, and is also
installed on Threat Grid Appliances. There is no communication between Threat Grid Cloud
service and the Threat Grid Portal that is included with a Threat Grid Appliance.
The Threat Grid Appliance v2.11 release updates the Threat Grid application to release
3.5.50.
Network Interfaces
The available network interfaces are described in the following table:
Interface
Description
Admin
Connect to the Admin network. Only inbound from Admin
network.
Admin UI traffic
SSH (inbound) for TGSH Dialog
NFSv4 for backups and clustering (Outbound. If a NFS hostname
is used instead of IP, this name will be resolved via Dirty
DNS.) Must be accessible from all cluster notes.
The Admin port can be disabled (from the tgsh shell); from
the Admin UI with v2.11. When disabled, non-clustered Threat
Grid Appliances can operate correctly with only the clean
and dirty ports connected, and the admin UI will be
presented on port 8443 of the clean interface (an also port
18443 with the v2.11 release). If the port is not disabled,
unplugging the admin port results in a non-functional (or at
best, a partially functional) Threat Grid Appliance.
Clust interface required for clustering (optional)
Requires an additional SFP+ module for direct interconnect.
This interface does not require any configuration. Addresses
are automatically assigned.
Clean
Connect to the Clean network. Clean must be accessible from
the corporate network but requires no outbound access to the
Internet.
UI and API traffic (inbound)
Sample submissions
SMTP (outbound connection to the configured mail server)
SSH (inbound for TGSH Dialog)
Syslog (outbound to configured syslog server)
ESA/WSA and CSA Integrations
AMP for Endpoints Private Cloud Integration
DNS optional
LDAP (outbound)
RADIUS (outbound)
NTP (for using an internal NTP server)
Dirty
Connect to the Dirty network; requires Internet access. Outbound
Only.
You should not use your own DNS (private IP) for the Dirty Interface
because traffic sent to a private IP is dropped at the Network Exit
Localization firewall.
DNS
Note
If you are setting up an integration with a AMP for
Endpoints Private Cloud, and the AMP for Endpoints
appliance hostname cannot be resolved over the Dirty
interface, then a separate DNS server that uses the
Clean interface can be configured in the Admin UI.
NTP (defaults to Dirty)
Updates
Support session in Normal operations mode
Support snapshots
Malware sample-initiated traffic
Recovery mode support session (outbound)
OpenDNS, TitaniumCloud, VirusTotal, ClamAV
SMTP outbound connections are redirected to a built-in
honeypot
Note
Using IPv4LL address space (168.254.0.16) for the Dirty interface
is not supported.
CIMC Interface
If the Cisco Integrated Management Controller (CIMC) interface is
configured, it can be used for server management and maintenance.
See CIMC
Configuration.
Note
CIMC is not supported on the Threat Grid M5 Appliance server.
Network Interface Setup Diagram
This section describes the most logical and recommended setup for a Threat Grid
Appliance. However, each customer's interface setup is different. Depending on your
network requirements, you may decide to connect the Dirty interface to the inside, or
the Clean interface to the outside with appropriate network security measures in place.
Note
In Threat Grid Appliance (v2.7.2 and later), the enable_clean_interface option
is available but is disabled by default. This option (after applying configuration
and rebooting) enables access to the administrative interface on port 8443 of the
assigned clean IP.
Firewall Rules
This section provides suggested firewall rules.
Note
Implementing a restrictive outgoing policy on the Dirty interface for ports 22 and
19791 requires tracking updates over time and spending more time maintaining the
firewall.
Note
Using IPv4LL address space (168.254.0.16) for the Dirty interface is not
supported.
Dirty Interface Outbound
Source
Destination
Protocol
Port
Action
Note
Dirty Interface
Internet
ANY
ANY
Allow
Allow outbound traffic from samples. (To get accurate results it
is required that malware be allowed to contact its command and
control server using whatever port and protocol it is designed to
use.)
Dirty Interface Inbound
Source
Destination
Protocol
Port
Action
Note
ANY
Dirty Internet
ANY
ANY
Deny
Deny all incoming connections.
Clean Interface Outbound
Source
Destination
Protocol
Port
Action
Note
Clean Interface
SMTP Servers
TCP
25
Allow
The appliance uses the clean interface to initiate SMTP
connections to the configured mail server.
Clean Interface Outbound (Optional)
Source
Destination
Protocol
Port
Action
Note
Clean Interface
Corporate DNS Server
TCP/UDP
53
Allow
Optional, only required if Clean DNS is configured.
Clean Interface
AMP Private Cloud
TCP
443
Allow
Optional, only required if AMP for Endpoints Private Cloud
integration is used.
Clean Interface
Syslog Servers
UDP
514
Allow
Allow connectivity to server designated to receive Syslog
messages and Threat Grid notifications.
Clean Interface
LDAP Servers
TCP/UDP
389
Allow
Optional, only required if LDAP is configured.
Clean Interface
LDAP Servers
TCP
636
Allow
Optional, only required if LDAP is configured.
Clean Interface
RADIUS Servers
DTLS
2083
Allow
Allow login to Threat Grid application UI (Face). Optional, only
required if RADIUS is configured.
Clean Interface
Internet
UDP
123
Allow
Optional, use this off-by-default functionality to use an
internal NTP server.
Clean Interface Inbound
Source
Destination
Protocol
Port
Action
Note
User Subnet
Clean Interface
TCP
22
Allow
Allow SSH connectivity to the TGSH Dialog.
User Subnet
Clean Interface
TCP
80
Allow
Appliance API and Threat Grid user interface. This will redirect
to HTTPS TCP/443.
User Subnet
Clean Interface
TCP
443
Allow
Appliance API and Threat Grid user interface.
User Subnet
Clean Interface
TCP
9443
Allow
Allow connectivity to the Threat Grid UI Glovebox.
Admin Interface Outbound (Optional)
The following depends on what services are configured.
Source
Destination
Protocol
Port
Action
Note
Admin Interface
NFSv4 Server
TCP
2049
Allow
Optional, only required if Threat Grid Appliance is configured to
send backups to an NFSv4 share.
Admin Interface Inbound
Source
Destination
Protocol
Port
Action
Note
Admin Subnet
Admin Interface
TCP
22
Allow
Allow SSH connectivity to the TGSH Dialog.
Admin Subnet
Admin Interface
TCP
80
Allow
Allow access to the Admin UI. This will redirect to HTTPS
TCP/443.
Admin Subnet
Admin Interface
TCP
443
Allow
Allow access to the Admin UI.
Dirty Interface for Non Cisco-Validated/Recommended Deployment
Source
Destination
Protocol
Port
Action
Note
Dirty Interface
Internet
TCP
22
Allow
Update, support snapshot, and licensing services.
Dirty Interface
Internet
TCP/UDP
53
Allow
Allow outbound DNS.
Dirty Interface
Internet
UDP
123
Allow
Allow outbound NTP.
Dirty Interface
Internet
TCP
19791
Allow
Allow connectivity to Threat Grid support.
Dirty Interface
Cisco Umbrella
TCP
443
Allow
Connect with third-party detection and enrichment
services.
Dirty Interface
VirusTotal
TCP
443
Allow
Connect with third-party detection and enrichment
services.
Dirty Interface
TitaniumCloud
TCP
443
Allow
Connect with third-party detection and enrichment
services.
Privacy and Sample Visibility
When submitting samples to a Threat Grid Appliance for analysis, an important
consideration is the privacy of the content. Privacy is a particularly important
consideration if sensitive documents or archive types are submitted for analysis,
because locating sensitive material could be relatively easy for those with access to
the Threat Grid Appliance, especially with the search API.
The privacy and sample visibility model for sample submissions to Threat Grid is as
follows:
Unless samples are designated as Private, they are visible to users who are
outside the submitter's organization.
Private samples can only be seen by Threat Grid users within the same
organization as the user who submitted the sample.
Samples Submitted by Integrations
The privacy and sample visibility model is modified on Threat Grid Appliances for samples
that are submitted by integrations. Integrations are Cisco products such as Email
Security Appliance (ESA), Web Security Appliance (WSA), and other devices or third-party
services (you may see the term CSA Integrations, which refers to ESA/WSA and other Cisco
appliances, devices, and services that are integrated; for example, registered, with
Threat Grid Appliance via the Cisco Sandbox API.)
All sample submissions on Threat Grid Appliances are Public by default, and can be viewed
by any other appliance user, including integrations, regardless of the organization to
which they belong. All appliance users can see all details of samples submitted by all
other users.
Threat Grid users may also submit Private samples to the Threat Grid Appliance, which are
only visible to other Threat Grid Appliance users, including integrations, from the same
organization as the sample submitter.
Privacy and sample visibility model on Threat Grid Appliances are illustrated in the
table.
Full Access - The green check mark indicates that users have full access
to the sample and the analysis results.
Scrubbed Reports - The grey check mark indicates that the Private
submission results are scrubbed. Users have partial access to the sample and
analysis results, but all potentially sensitive information about the sample is
removed. There are no filenames, process names, screenshots, or even specifics
about its activity in the glovebox.
We omit details from the Metadata section, such as the sample submitter's login
information. If you encounter a hash from a private sample in the course of
doing business, this will let alert you to known threats, and if you need more
details, submit your own copy of the sample for full analysis.
Private samples may not be downloaded. Scrubbed reports include Artifacts (with
filename removed), Behavioral Indicators, Domains, and IPs.
No Access - The red X indicates that users have no access to the sample or
the analysis results.
The same basic privacy rules apply to Threat Grid Appliance integrations with AMP for
Endpoints Private Cloud.
Wipe Appliance Operation
The Wipe Appliance operation enables you to wipe the disks on a Threat Grid Appliance to
remove all data prior to decommissioning or returning it to the Cisco Demo Loan
Program.
Important
After performing the wipe appliance procedure, the Threat Grid Appliance will no
longer operate without being returned to Cisco for reimaging.
Logs, active configuration, and other customer-owned data is now stored almost
exclusively on the RAID 5 data array, rather than being distributed between data
and OS drives. The remaining appliance-specific content stored on OS drives is
limited to information required for correct operation of recovery mode should
the data drives not be mountable, and has limited privacy impact if
disclosed.
Because less content is stored on the OS array with the
2.12 release, early appliances (with smaller OS drives) are less likely to need to
delete VM images other than the mandatory default image during a data reset (and
thus need to download updates online before those deleted VM images become available
again).