AWSCloudTrail - Successful brute force attack on S3 Bucket

Status: available
Severity: high
Time window: 1h
Group by: AccountName, AccountUPNSuffix, RecipientAccountId, SourceIpAddress, UserIdentityAccountId, bucketName
Source: github.com/Azure/Azure-Sentinel

Detects repeated failed GetObject attempts against an S3 bucket followed by a successful access from the same source, which can indicate brute-force-style object discovery or access attempts.

MITRE ATT&CK coverage

Tactic	Techniques
Credential Access	`T1110` Brute Force

Telemetry coverage

Platform	Record / event type
AWS	CloudTrail event `GetObject`: Retrieves an object from a specified S3 bucket.

Rules detecting the same action

These rules filter on the same operation.

Rule body

id: 31b9e94b-0df6-4a3d-a297-3457b53c5d86
name: AWSCloudTrail - Successful brute force attack on S3 Bucket
description: |
  Detects repeated failed GetObject attempts against an S3 bucket followed by a successful access from the
  same source, which can indicate brute-force-style object discovery or access attempts.
severity: High
status: Available
requiredDataConnectors:
  - connectorId: AWS
    dataTypes:
      - AWSCloudTrail
queryFrequency: 1h
queryPeriod: 1h
triggerOperator: gt
triggerThreshold: 0
tactics:
  - CredentialAccess
relevantTechniques:
  - T1110
query: |
    let timeframe = 1h;
    let failed_attempts = AWSCloudTrail
    | where TimeGenerated >= ago(timeframe)
    | where EventName == "GetObject" and isnotempty(ErrorMessage) and isnotempty(ErrorCode)
    | where UserIdentityAccountId == "ANONYMOUS_PRINCIPAL" or UserIdentityAccessKeyId <> RecipientAccountId
    | extend UserIdentityArn = iif(isempty(UserIdentityArn), tostring(parse_json(Resources)[0].ARN), UserIdentityArn)
    | extend UserName = tostring(split(UserIdentityArn, '/')[-1])
    | extend AccountName = case( UserIdentityPrincipalid == "Anonymous", "Anonymous", isempty(UserIdentityUserName), UserName, UserIdentityUserName)
    | extend AccountName = iif(AccountName contains "@", tostring(split(AccountName, '@', 0)[0]), AccountName),
      AccountUPNSuffix = iif(AccountName contains "@", tostring(split(AccountName, '@', 1)[0]), "")
    | extend bucketName = tostring(parse_json(RequestParameters).bucketName), keyName = tostring(parse_json(RequestParameters).key)
    | summarize time_min_failed=arg_min(TimeGenerated, *), failed_keys = dcount(keyName) by RecipientAccountId, AccountName, AccountUPNSuffix, UserIdentityAccountId, SourceIpAddress, bucketName
    | where failed_keys > 20;
    let success_attempts = AWSCloudTrail
    | where TimeGenerated >= ago(timeframe)
    | where EventName == "GetObject" and isempty(ErrorMessage) and isempty(ErrorCode)
    | where UserIdentityAccountId == "ANONYMOUS_PRINCIPAL" or UserIdentityAccessKeyId <> RecipientAccountId
    | extend UserIdentityArn = iif(isempty(UserIdentityArn), tostring(parse_json(Resources)[0].ARN), UserIdentityArn)
    | extend UserName = tostring(split(UserIdentityArn, '/')[-1])
    | extend AccountName = case( UserIdentityPrincipalid == "Anonymous", "Anonymous", isempty(UserIdentityUserName), UserName, UserIdentityUserName)
    | extend AccountName = iif(AccountName contains "@", tostring(split(AccountName, '@', 0)[0]), AccountName),
      AccountUPNSuffix = iif(AccountName contains "@", tostring(split(AccountName, '@', 1)[0]), "")
    | extend bucketName = tostring(parse_json(RequestParameters).bucketName), keyName = tostring(parse_json(RequestParameters).key)
    | summarize time_min_success=arg_min(TimeGenerated, *), success_keys = dcount(keyName) by RecipientAccountId, AccountName, AccountUPNSuffix, UserIdentityAccountId, SourceIpAddress, bucketName
    | where success_keys >= 1;
    failed_attempts
    | join kind=inner success_attempts on SourceIpAddress, RecipientAccountId, AccountName, AccountUPNSuffix, UserIdentityAccountId, bucketName
    | where time_min_success > time_min_failed
    | project-away keyName
entityMappings:
  - entityType: Account
    fieldMappings:
      - identifier: Name
        columnName: AccountName
      - identifier: UPNSuffix
        columnName: AccountUPNSuffix
      - identifier: CloudAppAccountId
        columnName: RecipientAccountId
  - entityType: IP
    fieldMappings:
      - identifier: Address
        columnName: SourceIpAddress
customDetails:
  bucketName: bucketName
  failed_keys: failed_keys
  success_keys: success_keys
  RecipientAccountId: RecipientAccountId
alertDetailsOverride:
  alertDisplayNameFormat: 'AWS S3 brute force pattern detected from {{SourceIpAddress}}'
  alertDescriptionFormat: 'Detected repeated failed and then successful S3 GetObject access for bucket {{bucketName}} in account {{RecipientAccountId}}.'
version: 1.0.2
kind: Scheduled

Stages and Predicates

Parameters

let timeframe = 1h;

let failed_attempts is inlined into the numbered stages below.

Let binding: `success_attempts` used in Stage 8

let success_attempts = AWSCloudTrail
| where TimeGenerated >= ago(timeframe)
| where EventName == "GetObject" and isempty(ErrorMessage) and isempty(ErrorCode)
| where UserIdentityAccountId == "ANONYMOUS_PRINCIPAL" or UserIdentityAccessKeyId <> RecipientAccountId
| extend UserIdentityArn = iif(isempty(UserIdentityArn), tostring(parse_json(Resources)[0].ARN), UserIdentityArn)
| extend UserName = tostring(split(UserIdentityArn, '/')[-1])
| extend AccountName = case( UserIdentityPrincipalid == "Anonymous", "Anonymous", isempty(UserIdentityUserName), UserName, UserIdentityUserName)
| extend AccountName = iif(AccountName contains "@", tostring(split(AccountName, '@', 0)[0]), AccountName),
  AccountUPNSuffix = iif(AccountName contains "@", tostring(split(AccountName, '@', 1)[0]), "")
| extend bucketName = tostring(parse_json(RequestParameters).bucketName), keyName = tostring(parse_json(RequestParameters).key)
| summarize time_min_success=arg_min(TimeGenerated, *), success_keys = dcount(keyName) by RecipientAccountId, AccountName, AccountUPNSuffix, UserIdentityAccountId, SourceIpAddress, bucketName
| where success_keys >= 1;

Stages 1 to 7 define let failed_attempts (the rule's main pipeline source); stages 8 to 10 run on it.

Stage 1: `source`

AWSCloudTrail

Stage 2: `where`

| where TimeGenerated >= ago(timeframe)

Stage 3: `where`

| where EventName == "GetObject" and isnotempty(ErrorMessage) and isnotempty(ErrorCode)

Stage 4: `where`

| where UserIdentityAccountId == "ANONYMOUS_PRINCIPAL" or UserIdentityAccessKeyId <> RecipientAccountId

Stage 5: `extend` (5 consecutive steps)

| extend UserIdentityArn = iif(isempty(UserIdentityArn), tostring(parse_json(Resources)[0].ARN), UserIdentityArn)
| extend UserName = tostring(split(UserIdentityArn, '/')[-1])
| extend AccountName = case( UserIdentityPrincipalid == "Anonymous", "Anonymous", isempty(UserIdentityUserName), UserName, UserIdentityUserName)
| extend AccountName = iif(AccountName contains "@", tostring(split(AccountName, '@', 0)[0]), AccountName),
  AccountUPNSuffix = iif(AccountName contains "@", tostring(split(AccountName, '@', 1)[0]), "")
| extend bucketName = tostring(parse_json(RequestParameters).bucketName), keyName = tostring(parse_json(RequestParameters).key)

Stage 6: `summarize`

| summarize time_min_failed=arg_min(TimeGenerated, *), failed_keys = dcount(keyName) by RecipientAccountId, AccountName, AccountUPNSuffix, UserIdentityAccountId, SourceIpAddress, bucketName

Stage 7: `where`

| where failed_keys > 20

Stage 8: `join`

failed_attempts
| join kind=inner success_attempts on SourceIpAddress, RecipientAccountId, AccountName, AccountUPNSuffix, UserIdentityAccountId, bucketName

Stage 9: `where`

| where time_min_success > time_min_failed

Stage 10: `project-away`

| project-away keyName

Indicators

These rows show field, operator, and value matches.

Field	Kind	Values	Search
`ErrorCode`	is_not_null	(no value, null check)	`field:"aws::errorCode" kind:is_not_null`
`ErrorCode`	is_null	(no value, null check)	`field:"aws::errorCode" kind:is_null`
`ErrorMessage`	is_not_null	(no value, null check)	`field:"aws::errorMessage" kind:is_not_null`
`ErrorMessage`	is_null	(no value, null check)	`field:"aws::errorMessage" kind:is_null`
`EventName`	eq	`GetObject`	`field:"aws::eventName" kind:eq value:"GetObject"`
`UserIdentityAccessKeyId`	cross_field_compare	`RecipientAccountId` transforms: `op:ne`	`field:"UserIdentityAccessKeyId" kind:cross_field_compare value:"RecipientAccountId"`
`UserIdentityAccountId`	eq	`ANONYMOUS_PRINCIPAL`	`field:"UserIdentityAccountId" kind:eq value:"ANONYMOUS_PRINCIPAL"`
`failed_keys`	gt	`20`	`field:"failed_keys" kind:gt value:"20"`
`success_keys`	ge	`1`	`field:"success_keys" kind:ge value:"1"`
`time_min_success`	cross_field_compare	`time_min_failed` transforms: `op:gt`	`field:"time_min_success" kind:cross_field_compare value:"time_min_failed"`

Output fields

These fields are emitted when the rule matches.

Field	Source
`AccountName`	`summarize`
`AccountUPNSuffix`	`summarize`
`RecipientAccountId`	`summarize`
`SourceIpAddress`	`summarize`
`UserIdentityAccountId`	`summarize`
`bucketName`	`summarize`
`failed_keys`	`summarize`
`time_min_failed`	`summarize`

AWSCloudTrail - Successful brute force attack on S3 Bucket

MITRE ATT&CK coverage

Telemetry coverage

Rules detecting the same action

Rule body

Stages and Predicates

Parameters

Let binding: success_attempts used in Stage 8

Stage 1: source

Stage 2: where

Stage 3: where

Stage 4: where

Stage 5: extend (5 consecutive steps)

Stage 6: summarize

Stage 7: where

Stage 8: join

Stage 9: where

Stage 10: project-away