GKE Endpoint Permission Enumeration

ELASTIC detection rule by Elastic · Elastic License v2

Detects a single authenticated GKE identity from one source IP issuing a burst of API calls across many distinct actions and resources with a mix of successful and failed outcomes. That pattern is consistent with automated RBAC permission enumeration rather than steady-state controller traffic. Anonymous probing is covered by a separate rule.

[metadata]
creation_date = "2026/07/21"
integration = ["gcp"]
maturity = "production"
updated_date = "2026/07/21"

[rule]
author = ["Elastic"]
description = """
Detects a single authenticated GKE identity from one source IP issuing a burst of API calls across many distinct
actions and resources with a mix of successful and failed outcomes. That pattern is consistent with automated RBAC
permission enumeration rather than steady-state controller traffic. Anonymous probing is covered by a separate rule.
"""
false_positives = [
    """
    Administrators troubleshooting RBAC from a workstation or VPN may generate mixed allow/deny results across many
    APIs in a short window; tune by identity or source IP after validation.
    """,
    """
    Newly deployed workloads with incomplete RoleBindings can probe several APIs during startup and encounter
    intermittent denials; exclude known service accounts if documented.
    """,
]
from = "now-6m"
interval = "5m"
language = "esql"
license = "Elastic License v2"
name = "GKE Endpoint Permission Enumeration"
note = """## Triage and analysis

### Investigating GKE Endpoint Permission Enumeration

The rule aggregates GKE audit events per `client.user.email` and `source.ip` over the rule lookback. It alerts when
the actor hits more than five distinct `event.action` values and more than three distinct `gcp.audit.resource_name`
values, produces both success and failure outcomes, and stays under 75 total events. Use
`Esql.earliest_timestamp` and `Esql.latest_timestamp` to bound the burst in Discover.

### Possible investigation steps

- Review `Esql.event_action_values`, `Esql.gcp_audit_resource_name_values`, and `Esql.event_outcome_values` for targeted APIs
  (secrets, RBAC, pods/exec) and which calls succeeded.
- Confirm whether `source.ip` and `Esql.user_agent_original_values` match expected admin or automation clients.
- Hunt for follow-on activity from the same identity: RoleBinding changes, secret reads, privileged pod creates, or
  exec.

### False positive analysis

- Platform engineers validating least-privilege RBAC can look like enumeration; correlate with change tickets.
- Chatty operators that occasionally fail authorization may approach the thresholds; raise exclusions only after
  confirming the identity is expected.

### Response and remediation

- If malicious, revoke or rotate the credential, tighten RBAC, and inspect for data access or persistence after the
  burst.
"""
setup = "The GCP Fleet integration with GKE audit logs enabled is required to be compatible with this rule."
references = [
    "https://heilancoos.github.io/research/2025/12/16/kubernetes.html#unauthenticated-api-access",
]
risk_score = 47
rule_id = "65985cd5-e654-40fe-8bc7-de26cb95309f"
severity = "medium"
tags = [
    "Domain: Cloud",
    "Domain: Kubernetes",
    "Data Source: GCP",
    "Data Source: Google Cloud Platform",
    "Use Case: Threat Detection",
    "Tactic: Discovery",
    "Resources: Investigation Guide",
]
timestamp_override = "event.ingested"
type = "esql"

query = '''
from logs-gcp.audit-* metadata _id, _index, _version
| where data_stream.dataset == "gcp.audit"
    and service.name == "k8s.io"
    and client.user.email is not null
    and source.ip is not null
    and to_string(source.ip) != "127.0.0.1"
    and to_string(source.ip) != "::1"
    and client.user.email != "system:anonymous"
    and client.user.email != "system:unauthenticated"
    and gcp.audit.resource_name != "readyz"
    and gcp.audit.resource_name != "livez"
    and gcp.audit.resource_name != "healthz"
    and gcp.audit.resource_name != "version"
| stats
    Esql.document_count = count(),
    Esql.event_outcome_count_distinct = count_distinct(event.outcome),
    Esql.event_action_count_distinct = count_distinct(event.action),
    Esql.gcp_audit_resource_name_count_distinct = count_distinct(gcp.audit.resource_name),
    Esql.earliest_timestamp = min(@timestamp),
    Esql.latest_timestamp = max(@timestamp),
    Esql.event_action_values = values(event.action),
    Esql.event_outcome_values = values(event.outcome),
    Esql.gcp_audit_resource_name_values = values(gcp.audit.resource_name),
    Esql.user_agent_original_values = values(user_agent.original)
  by client.user.email, source.ip
| where Esql.event_outcome_count_distinct == 2
    and Esql.event_action_count_distinct > 5
    and Esql.gcp_audit_resource_name_count_distinct > 3
    and Esql.document_count < 75
| keep Esql.*, client.user.email, source.ip
'''

[[rule.threat]]
framework = "MITRE ATT&CK"

[[rule.threat.technique]]
id = "T1613"
name = "Container and Resource Discovery"
reference = "https://attack.mitre.org/techniques/T1613/"

[rule.threat.tactic]
id = "TA0007"
name = "Discovery"
reference = "https://attack.mitre.org/tactics/TA0007/"

[[ currentRule.title && currentRule.title.length > 60 ? currentRule.title.slice(0, 60) + '…' : currentRule.title ]]

f21c577e-4dee-4a1b-8188-6fa4a56275ef

Propose edits, view related rules, and engage with the community through comments.

Executable code, not a declarative rule

This submission is arbitrary [[ currentRule.format ? currentRule.format.toUpperCase() : '' ]] code, not a declarative detection rule — review it as you would any script before running it, and note that it cannot be safely aggregated or auto-executed alongside YARA/Sigma/Suricata/CRS rules.

Rejected pattern present

[[ reason ]]

What does this mean?

Format & documentation suggestions

Not a security issue — this rule is valid and was accepted as-is. These are optional fields/conventions its own format recommends, mainly affecting its quality score and how well it documents itself for other reviewers.

[[ reason ]]

What do these messages mean?

[[ currentRule.title && currentRule.title.length > 60 ? currentRule.title.slice(0, 60) + '…' : currentRule.title ]]

[[ currentRule.title ]]

Community Discussion