Potential DNS Tunneling via NULL or Long DNS Queries

[metadata]
creation_date = "2026/08/20"
integration = ["network_traffic", "fortinet_fortigate", "zeek"]
maturity = "production"
updated_date = "2026/09/21"

[rule]
author = ["Elastic"]
description = """
Identifies DNS query patterns associated with tunneling, including rare NULL queries, repeated TXT queries with
unusually long names, and clients generating many unique, long subdomains under one registered domain within a
five-minute window.
"""
false_positives = [
    """
    CDN, cloud load-balancer, software-update, and telemetry hostnames can be long and change often. Recursive
    resolvers, forwarders, NAT gateways, and localhost DNS listeners can also combine queries from many endpoints under
    one client address. DNS research, testing, and security tools can generate NULL or unusually long TXT queries.
    Validate the apex domain and whether the source is an endpoint before treating the activity as tunneling.
    """,
]
from = "now-9m"
language = "esql"
license = "Elastic License v2"
name = "Potential DNS Tunneling via NULL or Long DNS Queries"
note = """## Triage and analysis

### Investigating Potential DNS Tunneling via NULL or Long DNS Queries

DNS tunneling encodes data in query labels and often produces many unique, unusually long subdomains under a single apex
domain. Some tunneling tools also use rare NULL queries or long TXT query names. This rule aggregates network DNS
telemetry for those behavioral patterns without relying on threat intelligence feeds or machine learning jobs.

Compare overlapping apex domains against the machine learning **DNS Tunneling** rule when that job is enabled.

### Possible investigation steps

- Review `Esql.dns_registered_domain`, `Esql.dns_question_type_values`, `Esql.total_qualifying_queries`,
  `Esql.count_distinct_names`, `Esql.unique_name_ratio`, `Esql.max_question_name_length`,
  `Esql.max_subdomain_length`, and sample values in `Esql.sample_names`.
- NULL alerts require at least three queries and two distinct names. Long-TXT alerts require at least five distinct
  names longer than 100 characters. The general long-subdomain branch retains the original threshold of 25 queries and
  15 distinct names across all query types.
- `Esql.count_queries`, `Esql.count_distinct_names`, and `Esql.unique_name_ratio` describe only long-subdomain
  activity. For alerts triggered only by NULL or long-TXT queries, both counts may be zero and the ratio will be null.
- TXT, NULL, CNAME, or MX bursts increase confidence; A/AAAA-only activity can still be tunneling and should not be
  dismissed on type alone.
- Use `Esql.first_seen`, `Esql.last_seen`, `Esql.dataset_values`, and `Esql.observer_name_values` to establish the event
  span and identify the integrations and sensors that contributed to the alert.
- Confirm whether `Esql.client_ip` is a workstation, server, recursive resolver, forwarder, NAT address, or localhost
  DNS service. Resolver and localhost sources merge many clients and are a common false-positive pattern.
- Review `Esql.destination_ip_values` to identify the resolver or authoritative destination observed by the sensor.
- Pivot on the same client and apex domain in raw DNS events and look for follow-on process, file, or additional C2
  activity.

### False positive analysis

- CDN, cloud load-balancer, certificate, and software-update services often create long hostnames. Confirm the apex
  domain reputation and whether the requesting host role normally uses that provider.
- Security or network appliances performing DNS-based reachability or reputation checks can resemble tunneling. Exclude
  confirmed appliance addresses after validation.
- Do not create a global resolver exception until the originating endpoint is known; a shared `Esql.client_ip` can hide
  a single infected host behind legitimate bulk lookups.

### Response and remediation

- Block the apex domain or forwarding from the affected host at recursive resolvers if malicious activity is confirmed.
- Isolate the source host and inspect for tunneling tools or malware initiating the queries.
- Add temporary blocks for the apex domain while scoping additional hosts querying the same name.
"""
references = [
    "https://code.kryo.se/iodine/",
    "https://unit42.paloaltonetworks.com/dns-tunneling-how-dns-can-be-abused-by-malicious-actors/",
]
risk_score = 47
rule_id = "89ed957d-609b-4b00-b8c6-a5cbd187632c"
setup = """## Setup

This rule requires DNS transaction events from one of the following sources:

- Elastic Network Packet Capture (`network_traffic.dns`) in `logs-network_traffic.dns-*`
- Fortinet FortiGate DNS logs (`fortinet_fortigate.log`) in `logs-fortinet_fortigate.log-*`
- Elastic Zeek (`zeek.dns`) in `logs-zeek.dns-*`
- Legacy Packetbeat DNS events in `packetbeat-*` with `event.dataset` set to `dns`

Place the sensor where it observes endpoint-to-resolver DNS traffic. If the sensor is upstream of a recursive resolver,
or if the captured client is a localhost listener such as `127.0.0.1`, `Esql.client_ip` may identify shared DNS
infrastructure instead of the originating endpoint.

DNS-over-HTTPS (DoH), DNS-over-TLS (DoT), and other encrypted DNS traffic are not visible to packet capture unless the
sensor receives decrypted DNS telemetry or equivalent resolver logs mapped to ECS.
"""
severity = "medium"
tags = [
    "Domain: Network",
    "Use Case: Threat Detection",
    "Use Case: Network Security Monitoring",
    "Rule Type: ES|QL",
    "Tactic: Command and Control",
    "Tactic: Exfiltration",
    "Data Source: Network Packet Capture",
    "Data Source: Fortinet",
    "Data Source: Network Traffic",
    "Data Source: Zeek",
    "Resources: Investigation Guide",
]
timestamp_override = "event.ingested"
type = "esql"

query = '''
from logs-network_traffic.dns-*, logs-fortinet_fortigate.log-*, logs-zeek.dns-*, packetbeat-*
| where
    (
      data_stream.dataset in ("network_traffic.dns", "fortinet_fortigate.log", "zeek.dns")
      or event.dataset == "dns"
    )
    and dns.question.name is not null
    and dns.question.registered_domain is not null
| eval
    Esql.client_ip = COALESCE(client.ip, source.ip),
    Esql.dataset = COALESCE(data_stream.dataset, event.dataset),
    Esql.dns_question_name = TO_LOWER(dns.question.name),
    Esql.dns_registered_domain = TO_LOWER(dns.question.registered_domain),
    Esql.dns_question_type = TO_LOWER(dns.question.type),
    Esql.question_name_length = LENGTH(Esql.dns_question_name),
    Esql.subdomain_length = LENGTH(Esql.dns_question_name) - LENGTH(Esql.dns_registered_domain) - 1
| where
    Esql.client_ip is not null
    and (Esql.dns_question_type is null or Esql.dns_question_type != "ptr")
    and not ENDS_WITH(Esql.dns_question_name, ".arpa")
    and (
      Esql.dns_question_type == "null"
      or Esql.subdomain_length >= 50
      or (Esql.dns_question_type == "txt" and Esql.question_name_length > 100)
    )
| eval
    Esql.long_subdomain_query_name = CASE(
      Esql.subdomain_length >= 50,
      Esql.dns_question_name,
      null
    ),
    Esql.null_query_name = CASE(
      Esql.dns_question_type == "null",
      Esql.dns_question_name,
      null
    ),
    Esql.long_txt_query_name = CASE(
      Esql.dns_question_type == "txt" and Esql.question_name_length > 100,
      Esql.dns_question_name,
      null
    )
| eval Esql.time_window = DATE_TRUNC(5 minutes, @timestamp)
| stats
    Esql.total_qualifying_queries = COUNT(*),
    Esql.count_queries = COUNT(Esql.long_subdomain_query_name),
    Esql.count_distinct_names = COUNT_DISTINCT(Esql.long_subdomain_query_name),
    Esql.null_query_count = COUNT(Esql.null_query_name),
    Esql.null_distinct_names = COUNT_DISTINCT(Esql.null_query_name),
    Esql.long_txt_query_count = COUNT(Esql.long_txt_query_name),
    Esql.long_txt_distinct_names = COUNT_DISTINCT(Esql.long_txt_query_name),
    Esql.max_question_name_length = MAX(Esql.question_name_length),
    Esql.avg_question_name_length = AVG(Esql.question_name_length),
    Esql.max_subdomain_length = MAX(Esql.subdomain_length),
    Esql.avg_subdomain_length = AVG(Esql.subdomain_length),
    Esql.dns_question_type_values = MV_SLICE(VALUES(Esql.dns_question_type), 0, 9),
    Esql.destination_ip_values = MV_SLICE(VALUES(destination.ip), 0, 4),
    Esql.sample_names = MV_SLICE(VALUES(Esql.dns_question_name), 0, 4),
    Esql.dataset_values = MV_SLICE(VALUES(Esql.dataset), 0, 9),
    Esql.observer_name_values = MV_SLICE(VALUES(observer.name), 0, 19),
    Esql.first_seen = MIN(@timestamp),
    Esql.last_seen = MAX(@timestamp)
  by Esql.time_window, Esql.client_ip, Esql.dns_registered_domain
| eval Esql.unique_name_ratio = CASE(
    Esql.count_queries > 0,
    TO_DOUBLE(Esql.count_distinct_names) / Esql.count_queries,
    null
  )
| where
    (
      Esql.count_queries >= 25
      and Esql.count_distinct_names >= 15
    )
    or (
      Esql.null_query_count >= 3
      and Esql.null_distinct_names >= 2
    )
    or (
      Esql.long_txt_query_count >= 5
      and Esql.long_txt_distinct_names >= 5
    )
| keep Esql.*
'''


[[rule.threat]]
framework = "MITRE ATT&CK"
[[rule.threat.technique]]
id = "T1071"
name = "Application Layer Protocol"
reference = "https://attack.mitre.org/techniques/T1071/"
[[rule.threat.technique.subtechnique]]
id = "T1071.004"
name = "DNS"
reference = "https://attack.mitre.org/techniques/T1071/004/"


[[rule.threat.technique]]
id = "T1572"
name = "Protocol Tunneling"
reference = "https://attack.mitre.org/techniques/T1572/"


[rule.threat.tactic]
id = "TA0011"
name = "Command and Control"
reference = "https://attack.mitre.org/tactics/TA0011/"
[[rule.threat]]
framework = "MITRE ATT&CK"
[[rule.threat.technique]]
id = "T1048"
name = "Exfiltration Over Alternative Protocol"
reference = "https://attack.mitre.org/techniques/T1048/"
[[rule.threat.technique.subtechnique]]
id = "T1048.003"
name = "Exfiltration Over Unencrypted Non-C2 Protocol"
reference = "https://attack.mitre.org/techniques/T1048/003/"



[rule.threat.tactic]
id = "TA0010"
name = "Exfiltration"
reference = "https://attack.mitre.org/tactics/TA0010/"

[rule.investigation_fields]
field_names = [
    "Esql.client_ip",
    "Esql.dns_registered_domain",
    "Esql.time_window",
    "Esql.total_qualifying_queries",
    "Esql.count_queries",
    "Esql.count_distinct_names",
    "Esql.unique_name_ratio",
    "Esql.null_query_count",
    "Esql.null_distinct_names",
    "Esql.long_txt_query_count",
    "Esql.long_txt_distinct_names",
    "Esql.dns_question_type_values",
    "Esql.max_question_name_length",
    "Esql.avg_question_name_length",
    "Esql.max_subdomain_length",
    "Esql.avg_subdomain_length",
    "Esql.destination_ip_values",
    "Esql.sample_names",
    "Esql.dataset_values",
    "Esql.observer_name_values",
    "Esql.first_seen",
    "Esql.last_seen",
]