↳ GitHub sourceAnalytics ruleMedium

Syntax errors stateful anomaly on database

Description

'This query batches of distinct SQL queries that failed with error codes that might indicate malicious attempts to gain illegitimate access to the data. When blind type of attacks are performed (such as SQL injection of fuzzying), the attempted queries are often malformed and fail on wrong syntax (error 102) or wrong escaping (error 105). Thus, if a large number of different queries fail on such errors in a short amount of time, this might indicate attempted attack.'
Rule type
Scheduled
Version
1.1.2
Declared status
Available
Query frequency
1h
Query period
14d
Trigger
gt 0

Declared MITRE coverage

Declared sources

Metadata from the source file. No dependencies inferred from KQL.

Connectors

Data types

KQL query

Original query, unchanged.

let monitoredStatementsThreshold = 1;           // Minimal number of monitored statements in the slice to trigger an anomaly.
let trainingSlicesThreshold = 5;                // The maximal amount of slices with monitored statements in the training window before anomaly detection is throttled.
let timeSliceSize = 1h;                         // The size of the single timeSlice for individual aggregation.
let detectionWindow = 1h;                       // The size of the recent detection window for detecting anomalies.  
let trainingWindow = detectionWindow + 14d;     // The size of the training window before the detection window for learning the normal state.
let monitoredErrors = pack_array(102, 105);     // List of sql error codes relevant for this detection.
let processedData = materialize (
    AzureDiagnostics
    | where TimeGenerated >= ago(trainingWindow)
    | where Category == 'SQLSecurityAuditEvents' and action_id_s has_any ("RCM", "BCM") // Keep only SQL affected rows
    | project TimeGenerated, PrincipalName = server_principal_name_s, ClientIp = client_ip_s, HostName = host_name_s, ResourceId,
              ApplicationName = application_name_s, ActionName = action_name_s, Database = strcat(LogicalServerName_s, '/', database_name_s),
              IsSuccess = succeeded_s, AffectedRows = affected_rows_d,
              ResponseRows = response_rows_d, Statement = statement_s,
              Error = case( additional_information_s has 'error_code', toint(extract("<error_code>([0-9.]+)", 1, additional_information_s))
                    , additional_information_s has 'failure_reason', toint(extract("<failure_reason>Err ([0-9.]+)", 1, additional_information_s))
                    , 0),
              State = case( additional_information_s has 'error_state', toint(extract("<error_state>([0-9.]+)", 1, additional_information_s))
                    , additional_information_s has 'failure_reason', toint(extract("<failure_reason>Err ([0-9.]+), Level ([0-9.]+)", 2, additional_information_s))
                    , 0),
              AdditionalInfo = additional_information_s, timeSlice = floor(TimeGenerated, timeSliceSize)
    | summarize countEvents = count(), countStatements = dcount(Statement), countStatementsWithError = dcountif(Statement, Error in (monitoredErrors))
        , anyMonitoredStatement = anyif(Statement, Error in (monitoredErrors)), anyInfo = anyif(AdditionalInfo, Error in (monitoredErrors))
        by Database, ClientIp, ApplicationName, PrincipalName, timeSlice,HostName,ResourceId
    | extend WindowType = case( timeSlice >= ago(detectionWindow), 'detection',
                                           (ago(trainingWindow) <= timeSlice and timeSlice < ago(detectionWindow)), 'training', 'other')
    | where WindowType in ('detection', 'training'));
let trainingSet =
    processedData
    | where WindowType == 'training'
    | summarize countSlicesWithErrors = dcountif(timeSlice, countStatementsWithError >= monitoredStatementsThreshold)
        by Database;
processedData
| where WindowType == 'detection' 
| join kind = inner (trainingSet) on Database
| extend IsErrorAnomalyOnStatement = iff(((countStatementsWithError >= monitoredStatementsThreshold) and (countSlicesWithErrors <= trainingSlicesThreshold)), true, false)
    , anomalyScore = round(countStatementsWithError/monitoredStatementsThreshold, 0)
| where IsErrorAnomalyOnStatement == 'true'
| project TimeGenerated = timeSlice, Database, ClientIp, ApplicationName, PrincipalName, HostName, ResourceId, countEvents, countStatements, countStatementsWithError, anyMonitoredStatement, anyInfo, anomalyScore 
| extend Name = tostring(split(PrincipalName,'@',0)[0]), UPNSuffix = tostring(split(PrincipalName,'@',1)[0])    

Declared entities

AccountIPHostCloudApplicationAzureResource

Related content

Links established from declared identifiers and solution manifests.

Source provenance

GitHub

Displayed values come from files in Azure/Azure-Sentinel. They describe the published template, not your workspace configuration.

Source identifier
c815008d-f4d1-4645-b13b-8b4bc188d5de
Additional source files 2Solutions/Azure SQL Database solution for sentinel/Analytic Rules/Detection-ErrorsSyntaxStatefulAnomalyOnDatabase.yamlsource ↗Solutions/Azure SQL Database solution for sentinel/Data/Solution_AzureSQLDatabasesolutionforsentinel.jsonsolution-membership ↗
GSTEP / CATALOG TRACKING

Added to catalog : 16 Sept 2026 · 05:49 UTC
Last change observed : 16 Sept 2026 · 05:49 UTC

GSTEP sync dates, separate from the source content’s publication dates.