Trace and alert on suspicious domains

Big picture

Add threat intelligence feeds to Calico Enterprise to trace DNS queries involving suspicious domains.

Value

Calico Enterprise integrates with threat intelligence feeds so you can detect when endpoints in your Kubernetes clusters query DNS for suspicious domains, or receive answers with suspicious domains. When events are detected, an anomaly detection dashboard in the UI shows the full context, including which pod(s) were involved so you can analyze and remediate.

Concepts

Pull or push threat feeds?

Calico Enterprise supports both push and pull methods for updating threat feeds. Use the pull method for fully automated threat feed updates without user intervention. Use the push method to schedule your own updates or if your threat feed is not available over HTTP(S).

Domain name threat feeds

A best practice is to develop an allow-list of "known-good" domains that particular applications or services must access, and then enforce this allow-list with network policy.

In addition to allow-lists, you can use threat feeds to monitor your cluster for DNS queries to known malicious or suspicious domain names. Calico Enterprise monitors DNS queries and generates alerts for any that are listed in your threat feed.

Threat feeds for domain names associated with malicious egress activity (e.g. command and control (C2) servers or data exfiltration), provide the most security value. Threat feeds that associate domain names with malicious ingress activity (e.g. port scans or IP sweeps) are less useful since these activities do not cause endpoints in your cluster to query DNS. It is better to consider IP-based threat feeds for ingress activity.

Before you begin...

Required

Privileges to manage GlobalThreatFeed.

Recommended

We recommend that you turn down the aggregation of DNS logs sent to Elasticsearch for configuring threat feeds. If you do not adjust DNS log aggregation settings, Calico Enterprise aggregates DNS queries from workloads in the same replica set. This means if a suspicious DNS query is detected, you will only know which replica set made the query and not which specific pod. Go to: FelixConfiguration and set the field, dnsLogsFileAggregationKind to 0 to log individual pods separately.

How to

This section describes how to pull or push threat feeds to Calico Enterprise.

• Pull threat feed updates
• Push threat feed updates

Pull threat feed updates

To add threat feeds to Calico Enterprise for automatic updates (default is once a day), the threat feed(s) must be available using HTTP(S), and return a newline-separated list of domain names.

Using Manager UI

1. From the Manager UI, select Threat Feeds --> Add Feed.
2. Add your threat feed on the Add a New Threat Feed window. For example:
   • Feed Name: feodo-tracker
   • Description: This is my threat feed based on domains.
   • URL: https://my.threatfeed.com/deny-list
   • Content type: DomainNameSet
   • Labels: Choose a label from the list.
3. Click Save Changes.
   From the **Action** menu, you can view or edit the details that you entered and can download the manifest file.

Go to the Alerts page to view events that are generated when an endpoint in the cluster queries a name on the list. For more information, see Manage alerts.

Using CLIs

1. Create the GlobalThreatFeed YAML and save it to file. The simplest example of this looks like the following. Replace the name and the URL with your feed.

   apiVersion: projectcalico.org/v3
   kind: GlobalThreatFeed
   metadata:
     name: my-threat-feed
   spec:
     content: DomainNameSet
     mode: Enabled
     description: 'This is my threat feed'
     feedType: Custom
     pull:
       http:
         url: https://my.threatfeed.com/deny-list

2. Add the global threat feed to the cluster.

   kubectl apply -f <your_threatfeed_filename>

Go to the Alerts page to view events that are generated when an endpoint in the cluster queries a name on the list. For more information, see Manage alerts.

Push threat feed updates

Use the push method if your threat feeds that are not in newline-delimited format, not available over HTTP, or if you prefer to push updates as they become available.

1. Create the GlobalThreatFeed YAML and save it to file. Replace the name field with your own name. The name is important in the later steps so make note of it.

   apiVersion: projectcalico.org/v3
   kind: GlobalThreatFeed
   metadata:
     name: my-threat-feed
   spec:
     content: DomainNameSet
     mode: Enabled
     description: 'This is my threat feed'
     feedType: Custom

2. Add the global threat feed to the cluster.

   kubectl apply -f <your_threatfeed_filename>

3. Configure or program your threat feed to write updates to Elasticsearch. This Elasticsearch document is written using the tigera_secure_ee_threatfeeds_domainnameset.\<cluster_name>. alias and must have the ID set to the name of the global threat feed object. The doc should have a single field called domains, containing a list of domain names. For example:

   PUT tigera_secure_ee_threatfeeds_domainnameset.cluster./_doc/my-threat-feed
   {
       "domains" : ["malicious.badstuff", "hacks.r.us"]
   }

   Note that to push data to ES, you'll need to configure a policy that allows that information to reach the ES cluster. See the Elasticsearch document APIs for how to create and update documents in Elasticsearch.

   If no alias exists in your Elasticsearch cluster, configure to write the Elastic document by specifying an index. Since the index name <tigera_secure_ee_threatfeeds_domainnameset.cluster.linseed-{now/s{yyyyMMdd}}-000001> contains a date pattern, make sure to send the request using the index name URL encoded.

   PUT /%3Ctigera_secure_ee_threatfeeds_domainnameset.cluster.linseed-%7Bnow%2Fs%7ByyyyMMdd%7D%7D-000001%3E/_doc/my-threat-feed
      {
          "domains" : ["malicious.badstuff", "hacks.r.us"]
      }

4. In Calico Enterprise Manager, go the “Alerts” page to view events that are generated when an endpoint in the cluster queries a name on the list.

Additional resources

See GlobalThreatFeed resource definition for all configuration options.