F5-specific SNMP settings are now available on the Polling Settings page, allowing you fine-tune the SNMP timeout, retry count, and max replies for F5 pollers. In addition, the default timeout value has been increased. These changes improve F5 polling reliability and reduce timeouts without affecting SNMP behavior for other devices.

Declarative Pipeline YAML parsing now correctly loads YAML definitions on systems using the Turkish locale by using an updated YAML parser, so configuration files are parsed without errors in this regional setting.

World Map management now saves manual map changes correctly on German-language installations, without showing SWIS query errors.

Hardware Health monitoring now keeps sensors visible when an unreachable device cannot be polled, changing their status to unknown instead of removing them from views, alerts, and reports.

Previously, unreachable nodes caused polling to silently report success with an empty result, which caused all hardware health sensor data to disappear from the UI, alerts, and reports. The advanced setting HardwareHealth.Pollers.HardwareHealth.SNMP.NPM.MaxSnmpTimeouts controls how many SNMP timeouts are tolerated before polling is considered failed. The default value is 3.

The advanced setting HardwareHealth.Data.HardwareHealth.IgnoredEventTypes is available to disable hardware health events generated by the collector during hardware health data processing. This can be used in environments with a large number of hardware health-capable nodes where limited connectivity results in SNMP request timeouts and causes many hardware health events to be generated. Accepted values are HardwareInfo, HardwareCategory, and HardwareItem. Separate multiple values with a comma.

An issue that caused the Configuration Wizard to fail with the following error was fixed:

Invalid object name 'Ctx_C_Orion_NetMan_CloudMonitoring_VirtualNetwork'

The WinRM/WMI SelectCommand for AppInsight now returns only the properties requested in a query instead of all properties, significantly lowering memory usage for large result sets such as IIS file statistics collection.

Alert integrations now ensure that when SolarWinds Observability Self-Hosted is integrated with ServiceNow, alert reset status is correctly synchronized to ServiceNow, and the SolarWinds Incident Response alert action works end-to-end so you can select an SIR instance, create incidents from alerts, and track their status updates directly in the SolarWinds Platform.

If an alert configuration becomes corrupted in the database, the alerting service no longer fails with the following error:

ERROR SolarWinds.Alerting.Service.AlertingService - Unhandled exception caught by Alerting Service.

An issue was resolved that caused the Configuration Wizard to fail during upgrades with the following message:

System.ServiceModel.EndpointNotFoundException: There was no endpoint listening

In Japanese and German environments, duplicated maps no longer show a different overall status than the original maps, even though they contain the same objects.

Network discovery deduplication controls now correctly honor the configuration that disables specific deduplication checks, allowing administrators to discover all intended nodes even when IP, MAC, or other deduplication rules are turned off.

Toolbar search fields now allow left and right arrow key movement inside the search box on affected pages, so you can move the text cursor with the keyboard as expected.

When a bulk update of Linux agents is performed and one or more agents does not have the .NET Framework installed, the update no longer fails.

Group and node status in the Web Console now stays consistent between the primary web server and additional web servers. Previously, the same object could show a different status (for example, Critical and OK) after the Configuration Wizard ran.

F5 LTM virtual server polling has been corrected so that the VirtualIPAddress / IPAddressIndex relationship is updated whenever the virtual IP (VIP) address changes, even when it was initially empty. This ensures that SolarWinds Observability Self-Hosted always uses the current VIP when correlating F5 virtual servers, which prevents missed or incorrect alerts when VIPs are reconfigured.

When an agent is assigned to an additional polling engine and automatic node creation is enabled, deploying the agent to the additional polling engine again after a failover to the primary polling engine no longer results in an error.

Discovery and report scheduling now correctly run advanced schedules from the specified Starting On date and time, instead of delaying the first run by an extra interval.

SD-WAN tunnel processing failures on Cisco Viptela Orchestrator devices were resolved. The issue was caused by duplicate tunnels in the device response (same tunnel, different destination port). Tunnel handling has been updated to include the port in the peer name, so duplicate entries no longer trigger the error An item with the same key has already been added, and all Viptela tunnel data is polled and displayed correctly.

You can now change the number of rows shown on the new All Active Alerts page. The page is no longer always limited to 20 results.

Discovery Ignore List now properly removes interfaces when you choose “Remove from Ignore List,” and those interfaces appear in scheduled discovery results again as expected.

Database Diagnostics were extended to include the new FortiGate firewall tunnel and virtual-domain tables or views, so diagnostics bundles now reliably capture configuration and metric data for FortiGate firewall devices without errors.

Node maintenance schedules now keep all node settings, such as credentials, intact when multiple schedules unmanage or remanage the same nodes at the same time. Previously, some settings could be cleared during these bulk actions.

On Arista switches with specific VLAN port configurations, VLAN-to-port mapping information was not being collected and displayed correctly. The issue has been resolved, and VLAN data is now accurately collected and reflected for affected devices.

Azure Virtual Network Gateways monitoring now continues to collect and store data even when some gateways do not have a Public IP address configured, preventing gaps in Azure gateway visibility caused by failed polling in these cases.

MIB Browser and Device Studio for SNMPv3 nodes now work correctly after upgrade, so you can browse and test OIDs on SNMPv3 devices without credential-related failures.

The Platform Connect advanced setting HideIconToAllUsers can now be used to hide the icon for all users without causing the Core Business Layer service to fail at startup.

When a user attempts to open the custom property editor from the legacy Manage Nodes view, the custom property editor now opens as expected instead of intermittently logging the user out and displaying the login dialog.

By default, the GET/POST alert action performs a header validation that can fail if the authorization header is not in the standard format. If this occurs, the centralized setting SendHttpRequestTokenHeaderValidation can be used to disable the header validation.

*After you edit a group from the Group Details page, clicking Save or Cancel no longer results in the following error:

The input is not a valid Base-64 string as it contains a non-base 64 character, more than two padding characters, or an illegal character among the padding characters.

*Azure SQL Database with Microsoft Entra ID (formerly Azure AD) authentication now allows the Feature Service and Business Layer to start correctly, so platform services can load and complete setup as expected.

*Special characters in a gMSA account name no longer prevent authentication.