The Job Engine no longer sends unnecessary messages to Linux Agents. The load generated by these messages could cause packet loss and connection issues.

In environments with High Availability pools, the SolarWinds Agent configuration was improved so that the Agents do not experience connection issues or do not connect to an incorrect polling engine. The issues were caused by incorrectly assigned engine IP addresses.

The SWQL query for unpaid HCO or module nodes has been optimized to prevent it from timing out in certain environments. The query now runs successfully and returns the correct number of licensed nodes.

AgentManagement message processing has been improved. This change fixes an issue that caused packet loss and high response times for agents, with messages such as Messaging.IncomingMessageProcessor - Unable to deliver response message.

When you select Root Level Groups as the view mode on the Manage Groups page, the pagination and the number of rows is correct.

When you run database maintenance manually, the swdebugMaintenanceExecutionHistory.log shows the time the job started as well as the time it completed.

On a computer with a Thai calendar, the configuration wizard no longer fails during an upgrade with the following messages:

Error running task Orion.MessageBus.MessageBusConfigurator Configuring message bus

SolarWinds.ConfigurationWizard.Common.ConfigurationException: Failed to setup RabbitMQ for Orion.

Opening a page containing PerfStack charts no longer causes the SolarWinds Platform to become unresponsive due to increased CPU and memory usage. In addition, the performance of PerfStack charts has been improved.

The macro ${N=Alerting;M=Notes} works correctly. When an alert definition includes that value, the associated information is included in the alert notification.

A condition that caused an extremely large number of alerts to be generated in some environments has been fixed.

SSH connections to a node from an additional web server no longer fail if a firewall is configured to restrict outbound connections to the node on port 22.

The SolarWinds Platform no longer fails periodically with errors like the following:

• There was an error retrieving data from SolarWinds Information Service

• Cannot access a disposed object. Object name: 'System.ServiceModel.Channels.ServiceChannel'.

If you have multiple alerts with the same condition, all of those alerts are triggered and reset as expected.

It is possible to add a node with more than 15,000 interfaces.

The SolarWinds Platform notifies you of failed login attempts in case an unknown user or a user with an invalid password is attempting to access the system. Invalid notifications are no longer sent when the login succeeds but the credentials were not in the default format.

To improve the performance of the Active Alerts view in the SolarWinds Web Console, alert history will be deleted from the SolarWinds Platform database after the specified retention period. By default, the retention period is 90 days. To change the default value:

1. Click Settings > All Settings.

2. Under Thresholds & Polling, click Polling Settings.

3. Change the value of the Alerts History Retention setting.

If your system is set to a language that uses commas to represent decimals, and you have custom properties that are floating point numbers with predefined values, it is now possible to enter or select values for these custom properties.

For security reasons, SolarWinds Platform 2023.2 added the dedicated database user ReportingClient with the Reporting role. This user queried the SolarWinds Platform database to gather report data. The user was granted only the privileges needed to perform that task. This user could not query other, custom databases; therefore, any reports, widgets, or queries to custom database stopped working.

To resolve this issue, use the Database Manager or the Server Management Studio to edit the WebSettings table in the SolarWinds Platform database. Change the value of ReportingSQLReportsUseDBServerLevelLogin to true.

When ReportingSQLReportsUseDBServerLevelLogin is set to true, the database account that was specified in the Configuration Wizard is used to query the database. By default, this is SolarWindsOrionDatabaseUser.

Request filtering rules that deny access to files with certain extensions can prevent the SolarWinds Platform installer and Web Console from functioning correctly. To prevent this issue, the Configuration Wizard now implements request filtering rules that explicitly allow access to the required file extensions.

To prevent timeouts when the passive agent protocol is sending data from the agent to AMS, throughput between the passive agent and the SolarWinds Platform server was improved by optimizing the transfer block size.

When multiple nodes have the same IP addresses, Network Discovery no longer adds interfaces to the wrong node.

The JobEngine agent correctly purges .result files.

This release fixes an issue that was preventing Network Atlas maps from opening or affecting performance for maps that did open.

In previous versions, HA tables needed to be truncated before a migration. If they weren't, a Centralized Upgrade would fail with the error System.Data.SqlClient.SqlException.

In 2023.4 and later, you no longer need to truncate HA tables before a migration, and duplicate rows in these tables no longer halt a Centralized Upgrade.

The Polling Engines page no longer lists the same additional polling engine multiple times in some circumstances.

The Unmanage Task Editor no longer adds an error message in the logs even though the task completed successfully.

The SolarWinds Installer no longer stops with the message that the Windows Event Log, Network List Service, and Network Location awareness applications must be closed for the installation to continue.

An attribute that could cause a memory leak has been removed from the SolarWinds Job Engine.

Alerts with an automatic reset condition no longer stop resetting after certain events, such as the database being temporarily disconnected.

Top XX Capacity widgets can be filtered using custom properties.

NFSv2 and NFSv3 network disks for AIX systems are now correctly classified as network disks instead of fixed disks.

A situation that could cause the Linux Job Engine to run a job outside its scheduled interval has been corrected.

The Job Engine no longer causes DNS lookups to fail.

If ServiceNow has an Authentication profile configuration, requests from the SolarWinds Platform to ServiceNow no longer fail with a 401 status code.^*

Volume polling no longer stops due to deadlocks occurring in the CoreThreshold table.^*

The throughput of the Job Engine service responsible for device polling was improved to fix issues related to polling completion below 100%, missed polls, or delayed polls.^*

The Configuration Wizard no longer fails when an empty database is selected.^*

In a large, complex environment, network configuration management jobs and inventory jobs run as expected.

Network configuration management jobs no longer make unnecessary APE license checks, which caused the jobs to run slowly.

Inventory collection was updated to prevent database blocks, which were affecting performance.

If you run a job that saves the results of a policy report to a file, and the file name includes a macro, the macro is parsed correctly and the report's content and formatting are accurate.

When SolarWinds high availability (HA) is deployed, a network configuration search performed after a failover includes all configs. The search is not limited to only the most recently downloaded configs.

A misspelled word in INFO messages in the NCM.Collector.Jobs_[*].log file has been corrected.

The Config Details page shows the downloaded time based on the SolarWinds Platform time zone, not UTC time.

When an inventory job retrieves a large Flash Size value, it records the value correctly and no longer returns the following error:

Arithmetic overflow error converting expression to data type int.

In 2023.2, the default timeout value for SWJobEngineWorker2x64.exe was increased from 20 minutes to 2 hours. In deployments with a large number of jobs that did not finish before the 2-hour timeout, memory usage increased significantly. To prevent this issue, the default timeout value has been decreased to 20 minutes.

When the SolarWinds Platform is configured to use Windows authentication, running interface percentile traffic reports no longer returns an error.

If an interface name includes special characters, the special characters are no longer escaped on the Interfaces with High Percent Usage (no paging) resource. This behavior was preventing database maintenance from running as scheduled.

The load time for the Interface Availability widget has been significantly decreased.

By default, network discovery enables all pollers on a device. So, if a topology poller is disabled on the List Resources or Manage Pollers page, it is enabled again when network discovery runs. A new advanced configuration option is available to enable you to override this default behavior. If do not want certain topology pollers to be enabled when network discovery runs:

1. Open the Advanced Configuration options page by pasting the following into your browser URL field after the hostname or IP address:

   /Orion/Admin/AdvancedConfiguration/Global.aspx

2. Under Topology, locate the PollersDiscoveryBlackList field.

3. Enter a comma-separated list of topology pollers that should not be enabled when network discovery runs.

Out-of-the-box Switch Stack alerts can be copied and edited, and they are triggered as expected.

If application data is corrupted, the Manage Applications and Service Ports page no longer attempts to list the applications. Corrupt data does not prevent the page from opening with the following message:

Unexpected Website Error
Sequence contains no elements

Additionally, a daily maintenance task detects and removes any corrupt application data.

If a user without permission to view NetFlow data attempts to access a NetFlow page, permissions are evaluated and the Restricted page message is displayed before the page loads. The page is no longer loaded and then hidden, and attempts to refresh the page do not affect performance.

If a deployment includes more than one application template with the same name, the Configuration Wizard no longer fails with the following error:

Database configuration failed: Error while executing script- Subquery returned more than 1 value.

The SolarWinds Platform Web Console no longer becomes unresponsive or stops functioning due to missing keys in the cloud monitoring resource file.

On the Node Details Summary page, when you click Real Time Event Viewer and then click the Message column for an event, the Message Details popup opens as expected.

When you add a node with Microsoft SQL Server 2022 deployed, AppInsight for SQL is discovered.

During upgrades data from old tables is migrated successfully and the obsolete tables are removed. In a previous version, incomplete data migration could cause issues such as:

• The Configuration Wizard failed with the message Invalid object name 'APM_ComponentStatus_Hourly.

• The SolarWinds Platform database size decreased significantly.

• The Application Component Details page did not consistently display messages.

If the Configuration Wizard fails during an upgrade, attempting to rerun the wizard no longer fails with the message Error while executing script- There is already an object named 'CLM_CloudJobSettings' in the database.

When a problem occurs with the API poller, the following message was improved to provide more information about the cause of the problem:

Index was outside the bounds of the array.

When a node is deleted or unmanaged, Application Dependency Mapping (ADM) polling does not continue for that node. In previous versions, continuing to poll a deleted or unmanaged node for ADM data caused the ADM_ProcessingQueue table to become very large.

Long status messages associated with an application monitor are displayed correctly after an upgrade.

Alerts on API Pollers that should be triggered when the response time exceeds a threshold are no longer triggered when the response time is below the threshold.

If ADM polling is disabled for the SolarWinds Platform, the node detail page does not display the Application Dependency Polling Enabled check box.

If AppInsight for IIS runs for long periods, the APM_IisBb_Request_Detail table no longer grows without limits, causing performance issues.

The default poll no longer fails when more than one array is in a Nimble cluster.

The Ethernet Ports Used Over Time now shows data for the selected time period instead of only the data for the current day.

When a wireless node is discovered with UDT port monitoring enabled and devices from the node are added to the Watchlist, the Watchlist widget no longer displays type conversion errors such as:

Conversion failed when converting the nvarchar value 'Vanguard_WLC' to data type int.

The Port Details widget now displays the MAC address, IP address, or Hostname for all devices.

The All User Log Ins widget displays data about the currently open endpoint, not the first endpoint that was opened.

When a node uses automatic private IP addressing (APIPA), the vendor and machine type are reported correctly.

The bulk insert query no longer runs slowly and causes performance issues on the database.

The Calls by Region widgets display the correct data.

IP SLA operations function correctly, and the ipsla.bussiness.layer log file no longer includes the message Violation of PRIMARY KEY constraint.

The FTP server is no longer filled with failed files, and VoIP views displays call manager information correctly.

After a call manager is removed, any associated CDR files are no longer processed, which prevents putting an unnecessary load on the FTP server.

Call managers are polled only by the assigned polling engine, not by all polling engines.

When CDR files for different call managers are stored in separate folders on the FTP server, CDR files for all call managers are processed. VNQM no longer ignores CDR files for all call managers except the last one added.

Avaya Call Manager now shows the correct numbers of phones.^*

Avaya RTCP data is collected and displayed as expected. Logs no longer include errors associated with the UTC offset.^*

Transactions no longer fail with the message Probe process has failed with exit code 2.