kql
421 TopicsMicrosoft Sentinel Query History not updating
Hello, Apologies if this isn't the correct place for this but I know I will likely retire before I get any traction with Microsoft support. Has anyone experienced issues with their Sentinel Query History not updating with the latest queries? I run a lot of queries each day and any time I open a new browser window and go to the logs tab, the latest query it shows in my history is 7/29/2025. If I run any new queries in that browser tab, they show in my query history but the moment I open a new browser tab and access sentinel logs, they are gone and it shows the latest query as 7/29/2025. My colleague has the exact same issue except their latest query date is 8/7/2025... Yes I do have the "Save query history" setting set to On. I have toggled it of and back on just to see if it would do anything but no luck. Does anyone know what could be causing this?179Views0likes6CommentsStandard Ontology and SIEM Field Mapping
Hello Community, We are working on a Microsoft Sentinel → Google Chronicle integration and need to automate the SIEM Field Mapping process between the two platforms Sentinel and google chronicle Schema Differences – Sentinel and Chronicle use different naming conventions and field hierarchies. Analytics Portability – Without mapping, a Chronicle rule expecting principal user email won’t understand Sentinel’s User Principal Name. Questions: Is there an API, PowerShell cmdlet, or Logic App method Sentinel’s field mapping with google chronicle fields.? is there any possibility via Automation.?71Views0likes2CommentsHow to exclude IPs & accounts from Analytic Rule, with Watchlist?
We are trying to filter out some false positives from a Analytic rule called "Service accounts performing RemotePS". Using automation rules still gives a lot of false mail notifications we don't want so we would like to try using a watchlist with the serviceaccounts and IP combination we want to exclude. Anyone knows where and what syntax we would need to exlude the items on the specific Watchlist? Query: let InteractiveTypes = pack_array( // Declare Interactive logon type names 'Interactive', 'CachedInteractive', 'Unlock', 'RemoteInteractive', 'CachedRemoteInteractive', 'CachedUnlock' ); let WhitelistedCmdlets = pack_array( // List of whitelisted commands that don't provide a lot of value 'prompt', 'Out-Default', 'out-lineoutput', 'format-default', 'Set-StrictMode', 'TabExpansion2' ); let WhitelistedAccounts = pack_array('FakeWhitelistedAccount'); // List of accounts that are known to perform this activity in the environment and can be ignored DeviceLogonEvents // Get all logon events... | where AccountName !in~ (WhitelistedAccounts) // ...where it is not a whitelisted account... | where ActionType == "LogonSuccess" // ...and the logon was successful... | where AccountName !contains "$" // ...and not a machine logon. | where AccountName !has "winrm va_" // WinRM will have pseudo account names that match this if there is an explicit permission for an admin to run the cmdlet, so assume it is good. | extend IsInteractive=(LogonType in (InteractiveTypes)) // Determine if the logon is interactive (True=1,False=0)... | summarize HasInteractiveLogon=max(IsInteractive) // ...then bucket and get the maximum interactive value (0 or 1)... by AccountName // ... by the AccountNames | where HasInteractiveLogon == 0 // ...and filter out all accounts that had an interactive logon. // At this point, we have a list of accounts that we believe to be service accounts // Now we need to find RemotePS sessions that were spawned by those accounts // Note that we look at all powershell cmdlets executed to form a 29-day baseline to evaluate the data on today | join kind=rightsemi ( // Start by dropping the account name and only tracking the... DeviceEvents // ... | where ActionType == 'PowerShellCommand' // ...PowerShell commands seen... | where InitiatingProcessFileName =~ 'wsmprovhost.exe' // ...whose parent was wsmprovhost.exe (RemotePS Server)... | extend AccountName = InitiatingProcessAccountName // ...and add an AccountName field so the join is easier ) on AccountName // At this point, we have all of the commands that were ran by service accounts | extend Command = tostring(extractjson('$.Command', tostring(AdditionalFields))) // Extract the actual PowerShell command that was executed | where Command !in (WhitelistedCmdlets) // Remove any values that match the whitelisted cmdlets | summarize (Timestamp, ReportId)=arg_max(TimeGenerated, ReportId), // Then group all of the cmdlets and calculate the min/max times of execution... make_set(Command, 100000), count(), min(TimeGenerated) by // ...as well as creating a list of cmdlets ran and the count.. AccountName, AccountDomain, DeviceName, DeviceId // ...and have the commonality be the account, DeviceName and DeviceId // At this point, we have machine-account pairs along with the list of commands run as well as the first/last time the commands were ran | order by AccountName asc // Order the final list by AccountName just to make it easier to go through | extend HostName = iff(DeviceName has '.', substring(DeviceName, 0, indexof(DeviceName, '.')), DeviceName) | extend DnsDomain = iff(DeviceName has '.', substring(DeviceName, indexof(DeviceName, '.') + 1), "")127Views0likes1CommentKQL: setting query time leads to problem in watchlist column projecting
Hello to the community! I have stumbled upon a very strange issue when using watchlists. I have a watchlist with 2 columns (userPrincipalName,allowedActivity) that I am then using to whitelist activities. Watchlist is imported using: let WhitelistedUsers = _GetWatchlist("testQuery") | project userPrincipalName, allowedActivity; Then I wanted to set it to a specific time frame to test it on given data set: set query_now = datetime("1/14/2022, 1:45:46.556 PM"); Problem is that when setting my query for a specific time, I get the following error from the watchlist: 'project' operator: Failed to resolve scalar expression named 'userPrincipalName'. Commenting the set query_now solves the project problem (not my problem though). I tried to set the time before and after watchlist import but that does not solve the issue. I could not find any posts around the topic (quite a specific one), so anyone observed similar behaviors or has a possible explanation? I can probably work around the set query_now with other functions but I gotten used to it, and find this behavior extremely strange4.3KViews0likes5CommentsIncident Missing Entities
Good morning! I would like to have some clarification on how entities work. Yesterday I found out that if I have 2 entities of the same type (In this particular case, two entities of the type Account), with the same identifier (originally, both share the identifier 'Name'), Sentinel appears to throw away one of them, or both in some instances, and when the alert generates an incident, the entities defined won't appear. I have switched out the identifier on both account types to something different, but until an incident gets triggered, I can't confirm if this will fix the original issue. So my questions are An analytic rule can or can't have two entities of the same type defined? If yes, that means that they need to have different identifiers. Is this a correct asumption? Some identifiers expect a certain type of value to be assigned, that means that eventually, I can ran out of identifiers for my entities or face the added complexity of dealing with types when returning values from my KQL query What could happen if I map an identifier to something that matches the type but not what that identifier represents? in this case, for Account, we have the identifier ObjectGuid. If I assign a value type string to it, that is not a guid, wouldn't that mess up something else in the background? Example, incident grouping If I move instead to use Sentinel Entities, which appear to be the 'general' option, I could only use one, since I only have 'entity' as available identifier, looping back to the problem of can have only one type of identifier for identity type. Thanks in advance103Views0likes2CommentsRecent Logic Apps Failures with Defender ATP Steps – "TimeGenerated" No Longer Recognized
Hi everyone, I’ve recently encountered an issue with Logic Apps failing on Defender ATP steps. Requests containing the TimeGenerated parameter no longer work—the column seems to be unrecognized. My code hasn’t changed at all, and the same queries run successfully in Defender 365’s Advanced Hunting. For example, this basic KQL query: DeviceLogonEvents | where TimeGenerated >= ago(30d) | where LogonType != "Local" | where DeviceName !contains ".fr" | where DeviceName !contains "shared-" | where DeviceName !contains "gdc-" | where DeviceName !contains "mon-" | distinct DeviceName Now throws the error: Failed to resolve column or scalar expression named 'TimeGenerated'. Fix semantic errors in your query. Removing TimeGenerated makes the query work again, but this isn’t a viable solution. Notably, the identical query still functions in Defender 365’s Advanced Hunting UI. This issue started affecting a Logic App that runs weekly—it worked on May 11th but failed on May 18th. Questions: Has there been a recent schema change or deprecation of TimeGenerated in Defender ATP's KQL for Logic Apps? Is there an alternative column or syntax we should use now? Are others experiencing this? Any insights or workarounds would be greatly appreciated!148Views1like3CommentsIdentityInfo with analytics KQL query
Hi, I'm currently trying to create a KQL query for an alert rule in Sentinel. The log source upon which the alert rule is based, only contains the SAMAccountName, which prevents me from mapping it to an Account entity in the alert. I'm therefore trying to use the IdentityInfo table to lookup the AadUserId of the user, using the SAMAccountName. The issue I'm running into is that I want my query to run every 10 minutes, and look up data from the past 10 minutes, as this is most suitable given the nature of the alert and the log source. This however causes the lookup in the IdentityInfo table to also only check data from the last 10 minutes, which doesn't work as the data in that table may be much older and therefor fail the lookup of the AadUserId of the user. According to the documentation, the IdentityInfo table is refreshed every 14 days, so for it to work I'd have to create a query that checks all logging, including that of the log source, from the past 14 days, which is not what I want. Hopefully some of you have suggestions or ideas on how to make this work. Thanks a lot! Marek331Views0likes8CommentsPlaybook when incident trigger is not working
Hi I want to create a playbook to automatically revoke session user when incident with specifics title or gravity is created. But after some test the playbook is'nt run autimacally, it work when I run it manually. I did'nt find what I do wrong. See the image and the code bellow. Thanks in advance! { "definition": { "$schema": "https://schema.management.azure.com/providers/Microsoft.Logic/schemas/2016-06-01/workflowdefinition.json#", "contentVersion": "1.0.0.0", "triggers": { "Microsoft_Sentinel_incident": { "type": "ApiConnectionWebhook", "inputs": { "host": { "connection": { "name": "@parameters('$connections')['azuresentinel']['connectionId']" } }, "body": { "callback_url": "@{listCallbackUrl()}" }, "path": "/incident-creation" } } }, "actions": { "Get_incident": { "type": "ApiConnection", "inputs": { "host": { "connection": { "name": "@parameters('$connections')['azuresentinel-1']['connectionId']" } }, "method": "post", "body": { "incidentArmId": "@triggerBody()?['object']?['id']" }, "path": "/Incidents" }, "runAfter": {} }, "Send_e-mail_(V2)": { "type": "ApiConnection", "inputs": { "host": { "connection": { "name": "@parameters('$connections')['office365']['connectionId']" } }, "method": "post", "body": { "To": "email address removed for privacy reasons", "Subject": "Ceci est un test", "Body": "</p> <p class="\"editor-paragraph\"">@{body('Get_incident')?['id']}</p> <p class="\"editor-paragraph\"">@{body('Get_incident')?['properties']?['description']}</p> <p class="\"editor-paragraph\"">@{body('Get_incident')?['properties']?['incidentNumber']}</p> <p>", "Importance": "Normal" }, "path": "/v2/Mail" }, "runAfter": { "Get_incident": [ "Succeeded" ] } } }, "outputs": {}, "parameters": { "$connections": { "type": "Object", "defaultValue": {} } } }, "parameters": { "$connections": { "type": "Object", "value": { "azuresentinel": { "id": "/subscriptions/xxxx/providers/Microsoft.Web/locations/xxxxx/managedApis/xxxxxxx", "connectionId": "/subscriptions/xxxxxxx/resourceGroups/xxxxxx/providers/Microsoft.Web/connections/azuresentinel-Revoke-RiskySessions1", "connectionName": "azuresentinel-Revoke-RiskySessions1", "connectionProperties": { "authentication": { "type": "ManagedServiceIdentity" } } }, "azuresentinel-1": { "id": "/subscriptions/xxxxxx/providers/Microsoft.Web/locations/xxxx/managedApis/xxx", "connectionId": "/subscriptions/xxxxxxx/resourceGroups/xxxxx/providers/Microsoft.Web/connections/xxxx", "connectionName": "xxxxxx", "connectionProperties": { "authentication": { "type": "ManagedServiceIdentity" } } }, "office365": { "id": "/subscriptions/xxxxxx/providers/Microsoft.Web/locations/xxxxx/managedApis/office365", "connectionId": "/subscriptions/xxxxx/resourceGroups/xxxxxx/providers/Microsoft.Web/connections/o365-Test_Send-email-incident-to-xxxx", "connectionName": "o365-Test_Send-email-incident-to-xxxxx" } } } } }Solved2.2KViews0likes2CommentsOptimisation For Abnormal Deny Rate for Source IP
Hi, I have recently enabled the "Abnormal Deny Rate for Source IP" alert in Microsoft Sentinel and found it to be quite noisy, generating a large number of alerts many of which do not appear to be actionable. I understand that adjusting the learning period is one way to reduce this noise. However, I am wondering if there are any other optimisation strategies available that do not involve simply changing the learning window. Has anyone had success with tuning this rule using: Threshold-based suppression (e.g. minimum deny count)? Source IP allowlists? Frequency filters (e.g. repeated anomalies over multiple intervals)? Combining with other signal types before generating alerts? Open to any suggestions, experiences, or best practices that others may have found effective in reducing false positives while still maintaining visibility into meaningful anomalies. Thanks in advance,181Views0likes1Comment