Here's what's new in SolarWinds Observability Self-Hosted 2026.4. You can find the applicable system requirements here.

Skipping a version? If you're upgrading from more than one version back, don't rely on this page alone — review the release notes aggregator to see everything that changed across the versions you’re skipping, and across your SolarWinds Platform products.

To view release notes, system requirements, and product guide (formerly known as administrator guide) PDFs for supported versions of SolarWinds Observability Self-Hosted, see SolarWinds Observability Self-Hosted previous versions.

 

Critical update advisories

AWS monitoring: Legacy CloudWatch cross-account monitoring was replaced with the cross-account IAM role (AssumeRole) method.

Azure Virtual Network Gateways were migrated to a new data polling framework. As a result, some entities, alerts, and centralized settings were removed and database settings were renamed. See Monitor several AWS accounts with IAM roles (AssumeRole).

Legacy wireless SWIS entities

Legacy wireless SWIS entities were removed in this release. As a consequence, custom alerts, reports, or widgets using the removed SWIS entities stop working. Replace the legacy SWIS entities with currently supported ones. See this topic for details.

SolarWinds Message Bus replaces RabbitMQ

All deployments that do not have High Availability configured will be upgraded to the new message bus, SolarWinds Message Bus, replacing RabbitMQ. In previous versions, this migration was limited to only deployments with fewer than 300 nodes, fewer than 3 polling engines, and not using High Availability.

Passwords for passive WPM Players

Included in this release are changes to Web Performance Monitor (WPM) components. All features and platform modules are installed regardless of active licensing, as all SolarWinds Platform products share an underlying framework. SolarWinds recommends upgrading all environments at your earliest convenience to apply the latest performance and security fixes.

Three types of WPM player installations are affected by this update advisory:

1. The default player installed on the main polling engine when SolarWinds Observability Self-Hosted was first installed.

2. Remote players installed by the customer and configured for passive mode.

3. Remote players installed by the customer in active mode.

After upgrading, passive WPM Players installed by default on the main polling engine (using Server-initiated communication) are switched to the active mode (using Player-initiated communication).

Passive WPM Players installed remotely are pre-filled with a randomly generated strong password during their automatic upgrade. This includes WPM Players with a disabled "Enable Upgrade" toggle. These are not upgraded to the latest version.

Active WPM Players (using Player-initiated communication) do not require a password to be set after the upgrade.

New features and improvements in SolarWinds Platform

SLA Mute mode for maintenance windows

Maintenance windows now support an SLA Mute state. In this state, monitoring and data collection continue, you can optionally mute alerting, and the SolarWinds Platform excludes the period from availability and SLA calculations. Unmanaging or disabling an object stops monitoring entirely, while SLA Mute keeps the object monitored during planned maintenance.

Redesigned network discovery

Network discovery has a redesigned workflow that uses the same interaction patterns as the rest of the modernized SolarWinds Platform Web Console and takes fewer steps to bring devices under monitoring.

Alert when a device stops sending SNMP traps

A new out-of-the-box alert triggers when a device does not send an SNMP trap within a time period that you specify. Use this alert to detect silent failures on devices that cannot be polled, such as a trap source behind a data diode.

Asset ID included in forwarded alerts

Alerts forwarded to integrated service management tools, including SolarWinds Service Desk, ServiceNow, and SolarWinds Incident Response, now include the Asset ID of the affected entity. The receiving tool can link the incident to the exact configuration item in its configuration management database (CMDB).

Accessibility improvements

Updated accessibility support for keyboard navigation, ARIA roles and semantics, and screen readers to help meet Section 508 requirements.

• HTML and ARIA accessibility: Corrected heading semantics, breadcrumb roles, and required parent roles for tables and menus. Added accessible names for overlays and switches, and corrected invalid or incomplete ARIA relationships.

• Keyboard accessibility: Enabled keyboard access to scrollable regions and search controls, improved the focus order in search, and removed focusable elements nested within other interactive controls.

• Search and main menu accessibility: Improved focus indicators and contrast, keyboard activation, Enter and Escape key behavior, accessible labels, and screen reader support for search controls.

• Interactive controls: Corrected nested interactive elements in radio controls and list items.

Scalability updates

A SolarWinds Platform Web Server now supports up to 150 concurrent user sessions, an increase from 100.

Concurrent users are counted as users interacting with the SolarWinds Platform Web Console within a 30-second window. Hardware requirements depend on the number of concurrent users supported.

FIPS 140-3 compliance

When deployed in FIPS-approved mode, SolarWinds Observability Self-Hosted 2026.4 and all listed modules are FIPS 140-3 ready, with documented exceptions.

Database Performance Analyzer Integration Module improvements

The following improvements have been added to the Database Performance Analyzer Integration Module (DPAIM):

• Cloud SQL databases monitored by DPA can be automatically discovered through DPAIM. Configuration changes are synchronized and obsolete cloud relationships are cleaned up automatically when cloud resources or accounts are deleted.

• From the DPA Server Management page, you can now edit the properties of an integrated DPA server without having to remove and then reestablish the integration.

New features and improvements in SolarWinds Observability Self-Hosted Essentials

The following features are exclusive to customers with a SolarWinds Observability Self-Hosted license.

SolarWinds AI assistant

At a future date, using AI assistant will consume credits based on the product tier and node count you are licensed for. Credit entitlements will be announced on or before credit enforcement starts.

Custom DHCP Options support

IPAM supports vendor-specific and site-specific DHCP options using private-use option codes 128–254. This capability is available in the DHCP Scopes area through the Manage Custom Options setting for users with the Power User role.

• Administrators can define custom option codes, data types, applicable protocol versions, supported server types, default values and help text.
• Declared options can be recognized during DHCP server scans and displayed alongside standard options in DHCP scopes.
• Users can add custom options when creating or editing scopes.

See The IPAM DHCP and DNS Management view and Define scope options on a DHCP server.

AWS VPC support

Extends IPAM capabilities to include Amazon Virtual Private Cloud (AWS VPC) and their associated subnets, and IP addresses of monitored VMs. A SAM, VMAN, or SolarWinds Observability Self-Hosted Advanced license is needed to use this functionality.

Advanced Search improvements

• Optimized search performance, including searches using Custom Properties and across additional object types.
• Improved search result pagination for more responsive navigation.
• Enhanced search results export with an improved interface.

Custom properties for LA rules

Log Analyzer rules are now a supported entity in Manage custom properties. See Add custom properties to LA rules.

Rule description for log rules

An optional rule description field is now available in LA log rules. Users can enter a description when creating a new rule and can add a description to existing rules when editing the rule.

Alert when no SNMP trap is received

LA now stores the time of the last trap received from each node. You can build an alert on this value to be notified when a node has been silent for a given amount of time. See Alert when a node stops sending SNMP traps.

Load Balancer F5 VELOS and R-Series support

Monitoring for F5 VELOS chassis and rSeries appliances now includes CPU and memory utilization, disk volumes, hardware health, machine type, and vendor details.

Extended Ruckus switch monitoring

Monitoring Ruckus switches now covers switch stack topology and health, hardware sensors such as power supplies, temperature, and fan status, and VLAN configuration.

API Poller enhancements

• Support added for bulk assignments.

• Support added for saving an API poller as a reusable template and assigning it to multiple nodes.

• Updated management page with dedicated tabs: one to view assigned API Pollers and another that shows available API Poller templates.

HPE StoreOnce application template

Monitor specific components on HPE StoreOnce devices using the new application monitor template.

Cisco CallManager (CCM) settings

Added new parameters for Cisco CallManager (CCM) phones and gateway polling in dbo.VoipConfig:

• PhoneConfirmationThresholdType (default: SuccessfulPollingSessions)
  • Values include: Hours, Days, SuccessfulPollingSessions
• PhoneConfirmationWithin (default: 2)
• GatewayConfirmationThresholdType (default: SuccessfulPollingSessions)
• GatewayConfirmationWithin (default: 2)

These settings replace the old day-based phone/gateway confirmation settings (PhoneDaysToLiveWithoutConfirmation / GatewayDaysToLiveWithoutConfirmation) with a more flexible threshold.

New features and improvements in SolarWinds Observability Self-Hosted Advanced

Customers with a SolarWinds Observability Self-Hosted Advanced license receive all Essentials updates, plus the following features exclusive to Advanced customers.

Traffic Key Analysis: Visualizing Network Flow Connections

Traffic Path Analysis introduces a Sankey-style visual flow diagram that shows how traffic moves across dimensions such as countries, endpoints, applications, protocols, and ports. Follow the width of each connection to identify the highest-volume paths and understand relationships that are difficult to see in separate rankings or pie charts. The feature includes predefined traffic-path views and a Custom view that lets you add, remove, and reorder dimensions.

• Predefined traffic views: Analyze geography by application or protocol, East-West and North-South traffic, protocol-port-application mismatches, unmonitored applications, IP Address Group usage, and application composition.
• Custom traffic paths: Build a focused investigation by selecting and reordering available dimensions to answer different traffic-flow questions without opening a separate report.
• High-volume path analysis: Use connection width to identify dominant traffic paths and trace traffic from its source through applications, protocols, or ports to its destination.

Manual peer assignment for ambiguous peer keepalive IP addresses

Cisco Nexus switches that form a Virtual Port Channel (vPC) pair identify their partner device using a peer keepalive IP address. NCM normally detects the paired device automatically by looking at the downloaded device configurations.

It's possible for multiple Nexus devices to use the same peer keepalive IP address (for example, when several devices share a management VRF with a reused addressing scheme). When this happens, NCM can no longer tell which device is the real partner. Instead of possibly displaying the wrong device, NCM marks the vPC as Ambiguous, and you can use Manual Assignment to specify the correct partner.

A manually assigned peer always takes priority over automatic detection. The assignment is saved, and you can change or remove it at any time.

Nutanix polling now uses the v4 API

Nutanix polling has been migrated to the v4 API.

Expanded VMWare support

Support for VMWare vSphere Client 9.x and vCenter Server 9.x has been added.

New features and improvements in SolarWinds Observability Self-Hosted Premier

Updated polling API now available in Pure Storage/Everpure devices

SRM now features an updated polling API in Pure Storage/Everpure devices to use the most recent version, 2.x API, from Pure Storage.

Fixed CVEs

At SolarWinds, we prioritize the swift resolution of CVEs to ensure the security and integrity of our software. In this release, we have successfully addressed the following CVEs.

Fixed customer issues

Case number	Description	Platform product
01453211	When multiple Nexus devices have the same peer IP address, you can now manually assign a Nexus vPC peer keep-alive link to ensure the assignment is correct.	NCM
01744792	When a user specifies an empty database during database configuration, the Configuration Wizard no longer checks the database branding version, and the following message is not displayed: An exception occurred when fetching the DB branding version.	Platform
00640947, 01051755, 01076777, 01623669, 01680694, 01687213, 01965097, 02061571	Interface Alias column length has increased to 255 characters, allowing longer interface descriptions to be properly collected and displayed without truncation.	NPM
01819306, 02013881	When a Linux or Unix Script Component with multiple statistics is assigned to an application, the Application Details view now shows a distinct message for each statistic.	SAM
00319360, 00427016, 00507187, 00762394, 01039752, 01626993, 01925972, 02175470	When the option 'Operationally down' is selected in Network Discovery, administratively shut down interfaces are also imported by default. The setting Discovery-SkipInterfacePortByAdministrativeStatusOnImports is now available to change the default behavior. When this setting is enabled, administratively shut down interfaces are imported only when the 'Administratively shutdown' option is selected. The new setting can be enabled by issuing a command to the database: update Settings set CurrentValue = 1 where SettingID = 'Discovery-SkipInterfacePortByAdministrativeStatusOnImports'	NPM
01683539	Status icons load correctly for all supported interface types.	NPM
01791596	On the Edit Account page, the SecurityTab Menu Bar is now available and can be assigned to views.	Platform
01791596, 01802623, 01976085, 02076730	During upgrades, the system check now determines whether the WebSocket protocol for IIS is enabled, and displays a message if it is not. Intelligent Maps use this protocol to efficiently display real-time changes. If this protocol is disabled, live updates work, but performance is slower. The message includes a link to the Support article How to Enable the WebSocket Protocol for the SolarWinds Platform.	Platform
00952585, 01025487, 01676765, 01692130, 01759425, 01972362, 01984154, 02052155, 02135512, 02151844	Performance related to processing wireless collector data has been improved. In previous versions, wireless collector data processing could lead to database maintenance errors, high memory usage, or polling issues.	NPM
01767579	When the option to log in automatically is selected on the Manage EOC Sites page and the authentication type is Pass-through Authentication, testing the connection to a new site no longer fails.	EOC
02059398	Transactions no longer fail because of insufficient permissions for domain accounts on the SEUM log folder.	WPM
01692130, 01776286, 01899647, 01976398	To improve web performance and system load times, nodes-related data operations use the db.entities framework efficiently, reducing blocking, excessive reads, and slow updates across partitioned tables.	Platform
01892652	When database connectivity is lost, the alerting engine remains stable and does not stop and restart repeatedly across the primary and secondary HA servers.	Platform
01308295, 01965000, 02126403	The Add Node wizard validates vCenter discovery results and reports an error when the vCenter cannot be identified. The wizard no longer reports success when the VCenter was not added.	VMAN
01972784	The Mandatory Notice appears only once during SAML authentication and subsequent dashboard access.	Platform
01634745, 01940097, 02068093	Cisco Wireless Controller polling now supports the CISCO-LWAPP-WLAN-MIB in addition to the AIRESPACE-WIRELESS-MIB. Admins can configure a preferred MIB globally or per node, allowing the system to automatically fallback and enrich wireless data for access points, interfaces, clients, and SSIDs.	NPM
01968777	The load time and performance of the SQL Performance Counter details page have been improved.	SAM
02028380	When an agent-monitored node is moved to a different polling engine, the node remains up and polling continues.	Platform
02119564	HA pool setup correctly treats the Virtual IP address and Hostname as optional. If a Virtual IP address is entered and then cleared, it is no longer cached.	Platform
02036022	The Status Summary resource, which was intended for EOC deployments, can no longer be added to views in other SolarWinds Observability Self-Hosted deployments.	Platform
02039738	Custom properties named FailoverStatus or SyncStatus no longer prevent F5 resources from opening or cause the Configuration Wizard to fail with an Ambiguous column name error.	NPM
02035418	Discovery now detects Power Conversion Units (PCUs) with newer firmware that uses a different date format, and upgrading an existing node to the newer firmware does not affect polling.	Platform
02141186	A change to the default authentication method enables the APM plugin to poll MS SQL databases in a Security Technical Implementation Guide (STIG) environment.	SAM
02020536	When the database includes large tables, database maintenance no longer fails with the following error: Failed to execute 'Index Defragmentation [Database]' System.OverflowException: Value was either too large or too small for an Int32.	SAM
02045443	Manage Nodes, Alerts, and Reports no longer display duplicate entries for Power Control Units (PCUs) or UPS devices, ensuring accurate data organization and clean user interface filtering.	Platform
02038245	List resources jobs on SolarWinds Agent nodes now provide traceable life cycle logging across creation, scheduling, propagation, and execution stages.	Platform
02046127, 02117091	Metrics are correctly collected for Azure Express Route Virtual Network Gateways.	NPM
02037173	When a report with a comma in the title is exported to Excel using the Export to Excel button, the file name is no longer truncated.	Platform
02015382	Disk data for AWS-based VMs is collected, and the Storage column on the AWS Virtual Machines page displays the correct values instead of Unknown.	SAM
02020772	The Discovery Results page now correctly displays the status as "Imported" for Windows devices rather than an incorrect "Changed" status.	SCM
02173415	As you drill down through entities in the Modern Dashboards Entity Drilldown resource, breadcrumbs display information about previous selections and can be used for navigation.	Platform
02056831, 02110991	Hardware Health excludes non-existent HP RAID controllers with device ID 0 and empty serial numbers, preventing undefined sensor statuses and alerts.	SAM
N/A	The Force-enable multi-subnet failover setting is applied to the database connection string for all supported SQL Server authentication methods.	Platform
02083215	When you set up a one-time scheduled maintenance job, the interval is scheduled correctly even if the client and server are in different time zones.	Platform
02106646	When a device's interface subtype is 5 (for example, Aruba devices), the device can now be polled by API and flow data is available. A new centralized setting, AllowedInterfaceSubTypes, currently recognizes the subtypes 1, 2, and 5. This setting can be updated to recognize additional types.	NTA
02160609	On the Azure monitoring tabs, filtering by account name no longer filters out entities incorrectly.	SAM
02069681, 02135654, 02134762, 02142137	Ruckus SmartZone polling handles null uplink and downlink rates without errors and continues displaying wireless data.	NPM
02135171, 02161717	In the Add Node or Edit Node wizard, if the gMSA Account option is selected for a node that uses WMI polling, the Connection Mode automatically changes to WinRMOnly, and the Authentication Mechanism changes to Kerberos. These changes enable successful gMSA authentication. If the gMSA Account option is later deselected, these options revert to their previous values.	Platform
02093182, 02099195, 02104654, 02165764, 02168101, 02176177	Polling Completion displays 100% for new installations and accurately aggregates job engine completion values after polling activity begins.	Platform
N/A	Job Engine v3 reliably processes job requests even when request-body data arrives slowly.	Platform
02081056	Database maintenance removes orphaned data across GCP services, preventing stale cloud entities from consuming license capacity.	Platform
02101101	The Node Details Poller has been improved to reliably poll certain OIDs, allowing the associated devices to be added.	Platform
02025395	When an anomaly-based alert is triggered and a user adds notes to the triggered alert instance, the alert entity details page displays the correct data, not There is no data to show.	Platform
02054679	The All IP Addresses on Node resource correctly displays both IPv4 and IPv6 address information for SNMP nodes.	Platform
02109883	HA pools remain healthy and standby APEs start correctly after upgrades without duplicate engine records or manual database cleanup.	Platform
N/A	Audit events for moving an exact node into or out of Load Balancing display the node caption as a link to its details.	Platform
02132955, 00839577	Audit events accurately record an alert’s current enabled or disabled state after each change.	Platform
N/A	SWQL alert enable and disable actions are recorded in audit logs as Alert Enabled or Alert Disabled events.	Platform
02113768	The AD test correctly validates wildcard certificates assigned to website SSL bindings.	Platform
02143592	Custom property account limitations applied to IPAM hierarchy groups correctly control visibility without requiring the same property on IP Networks.	IPAM
02097908, 02102350	The Orion.NetObjectDowntime table accurately reflects nodes whose current status is Up.	Platform
02115776	Proxmox clusters display the correct status even when hostnames are reused across clusters.	VMAN
02102003, 02098769	When Linux nodes are polled with the SolarWinds Agent using SNMPv2c or SNMPv3, Asset Inventory retrieves and displays all results.	Platform
N/A	HA pool actions remain available after canceling the force-failover dialog.	Platform
02070584, 02141068, 02153394	The MIB database installer can be run from the Windows user interface. Starting the installer from the command line is no longer required.	NPM
01484292, 02114456	When new agents are added, the resulting notification accurately reports the number of agents that were added.	Platform
02084409, 02101320, 02141125	The Meraki API client follows HTTP 308 redirects to shard-specific cloud endpoints, allowing onboarding and polling to complete successfully.	NPM
02069085	Exporting a report to a PDF no longer fails when an alternative PDF engine is used.	Platform
02118385	When a custom certificate cannot be used for the SWIS REST API endpoint (for example, because of a missing or inaccessible private key), the issue no longer prevents SWIS from starting. The issue is logged, and the SolarWinds certificate is used instead.	Platform
02073197	The MS Graph SMTP client configuration provides a link to a Support article with configuration information.	Platform
02071129	OSPF routing neighbor polling now automatically ignores invalid 0.0.0.0 router IDs to prevent duplicate database entries and inaccurate alert statuses.	NPM
02055084, 02115513, 02151844	Administrators can override the collector BatchCountLimit for Hardware Health and SwitchStack polling to keep collector queues manageable. The centralized settings BatchCountLimit and BatchExpirationMilliseconds are now available.	NPM
02120230	An issue that caused the Job Engine to truncate data polling to stop was resolved.	Platform
02116585	Proxmox VE polling no longer fails when the client has no iSCSI (Internet Small Computer System Interface) storage.	VMAN
02116585	Proxmox VE virtual machines use the correct IP addresses, enabling accurate VM-to-node mapping.	VMAN
02120893	The first time an imported compliance report runs, it no longer inaccurately reports rule violations, requiring editing the rule and rerunning the report.	NCM
02109761, 02125559, 02130782	When a deployment does not include a license for NCM, the Configuration Wizard no longer creates the NCM Nodes view. When the NCM Nodes view was created without the NCM BusinessLayer to keep it synchronized, the view could become corrupt and cause polling and other features to fail.	Platform
02034433	PerfStack projects now maintain unique metric-entity pairs when multiple entities are added or project data is loaded.	Platform
02104938, 02136724, 02130258, 02132847	Values in the Weight column of the CPUMultiLoad_CS_cur table are no longer saved as 0, which prevents database maintenance from failing with divide-by-zero errors.	Platform
02129893	The Next Run column in the Upcoming Maintenance Schedules resource no longer displays Invalid Date if the browser's language is not EN-US.	Platform
02130797	Unassigning an Application Monitor Template from a large group of nodes no longer fails or appears to hang.	SAM
02129334	When the ImportCannedReports task fails, the Configuration Wizard provides detailed information about the issue.	Platform
02126753	When the Text > Content dialog is opened from a node on an Intelligent Map, custom properties with names that include id are no longer missing from the dialog.	Platform
02132078	Database connections use a default connect timeout of 120 seconds, which prevents issues with the Configuration Wizard.	Platform
02112185	Nodes with SAM AppInsight applications monitored via an agent no longer flip from unmanaged to managed and back during scheduled maintenance periods, allowing alerts to fire.	SAM
02114659	An issue was fixed that caused Intelligent Maps to display incorrect node statuses.	Platform
02127763, 02134145, 02136803	Proxmox VE 9.x clusters are added successfully through the VMAN Add Virtualization wizard and appear in the VMAN inventory.	VMAN
02130804	The Windows DHCP scanner can now scan an IPv6 scope that has over 580 leases without looping, exhausting memory, or inaccurately marking scopes as down.	IPAM
02055084, 02115513	When an entity is missing the DataRingRedundant property, SwitchStack processing no longer fails with an ArgumentOutOfRangeException, leaving polled data uncommitted. A default DataRingRedundant property is assigned, and processing continues.	NPM
02141442	VMAN documentation has been updated to reflect the permissions required for WMI polling with a non-administrator domain account.	VMAN
02128138, 02144677	Dell warranty polling successfully accepts client IDs in UUID format and populates warranty data for Dell nodes.	SAM
02155772	When a Linux node is added via SNMP and an application with Linux script monitor is assigned to it, the Components resource on the Application Details page no longer displays duplicated text.	SAM
02132721	SolarWinds Deployment Health correctly distinguishes Azure SQL Database from Azure SQL Managed Instance when evaluating database time zones.	Platform
02145372	The machine type for APC Automatic Transfer Switch 4G devices is no longer incorrectly identified as Smart-UPS 9605.	Platform
01967985, 02110438, 00924965, 01084552	Memory is reported accurately for Linux devices with available valid extended 64-bit memory OIDs.	Platform
02139334	The Configuration Wizard no longer fails to upgrade with the following error because an incorrect database version was reported: Invalid object name 'dbo.NCM_VulnerabilitiesAnnouncementsNodes'.	Platform
02118726, 02106120, 02139711, 02144082	The centralized setting SkipNotification was added to enable a periodic reconciliation timer that ensures notifications of scheduled jobs are not missed because of a temporary message bus outage or connection interruption. This setting is off by default. If needed, contact Support to enable the setting. This fix is also included in the 2026.2.3 release.	Platform
N/A	The out-of-the-box Duplex Mismatch alert more accurately identifies interfaces with mismatched duplex modes.	NPM
02105639	The NCM Database and Archive Maintenance job now removes orphaned EngineID entries from the database. These orphaned entries could cause timeouts and errors on NCM settings pages.	NCM
02144215, 02158471	Azure Cosmos DB database maintenance completes successfully when data is moved to the next gradation level.	Platform
02131498	SWJobEngineWorker2 no longer repeatedly fails due to memory exhaustion and then restarts. This issue interrupted in-process jobs and returned unknown component states.	SAM
02162083	WMI connections now use an increased connection timeout with a separate configurable query timeout.	Platform
02120396	An unsupported Windows time zone ID in a scheduled job no longer prevents the Core Business Layer from starting.	Platform
02075114	VNQM downloads and processes all CDR/CMR files from Call Managers without extensive delays due to incorrect base-name execution.	VNQM
02138099	When installing a scalability engine fails because the product catalog is corrupted or could not be downloaded, the error message provides a link to a Support article with additional information.	Platform
02151490	When gMSA credentials are stored in UPN format, real-time services (Real-Time Process Explorer, Service Control Manager, or Real-Time Event Log Viewer) no longer fail to open with the following error message: An exception occurred during polling processes. Error: Index was outside the bounds of the array.	SAM
02123605	OpenTlsConnection accepts valid SQL data source formats with a space before the port.	Platform
02159012	IPAM authorization no longer fails for users whose account ID contains LIKE wildcard characters (such as [, ], %, or _).	IPAM
02142112	Meraki MX appliance VLAN and interface IP addresses are correctly imported and indexed, and the NetFlow receiver does not treat them as unmanaged devices and discard the flows.	NPM
01970925, 02159107	Administrators can edit and delete Nutanix credentials from the VIM Settings page.	VMAN
02143825	SAM applications assigned through dynamic group templates are automatically removed when a node leaves the dynamic group, even when that node is also an ESXi host in a VMware VIM module.	SAM
02164520	After the polling method for a node has been changed, polling errors from the previous polling method no longer persist and incorrectly affect the node's status.	Platform
02163303	When a group includes a node with hardware health data and you add the Hardware Health Chart resource to the group details page, the page no longer fails to open with the following error: Unexpected Website Error Resource Key 'HHLIBCODE_DOY_10' was not found in Resource File 'HardwareHealth.Strings'	SAM
02159858	The Link to Thin AP Device view and Link to Autonomous AP Device view no longer cross-link an AP to the wrong counterpart when one AP's IP address is a substring of another IP address. Exact IP matching is now used.	NPM
02158968	Date/Time custom properties save dates with the correct year, including years 2030 and later.	Platform
02158849, 02159896, 02163696, 02165947, 02166951, 02168049, 02169367, 02172661, 02173782, 02180056	Time synchronization issues between the SolarWinds Observability Self-Hosted server and a remote computer connected to the server do not prevent users from unmanaging or muting a node from the Manage Nodes page. This fix is also included in the 2026.2.3 release.	Platform
02162683	GCP load balancer implementations report Private Service Connect connection details and global-access status correctly.	Essentials
02166138, 02164851	The Configuration Wizard removes deprecated VSAN functions from the database, preventing issues during database backup and maintenance operations.	NPM
02116254	Cisco ASA Site-to-Site VPN pollers successfully collect and display tunnel information when remote addresses use dotted-format notation.	NPM
02151105	GKE Zonal clusters are discovered and polled alongside Regional clusters when GKE monitoring is enabled for all regions.	Essentials
02154649, 02168426, 02169475, 02171488	SWIS queries correctly apply filters on node custom properties, including global, dashboard, and URL filters.	Platform
N/A	*Modern dashboard widgets return data for SWQL queries that include a FILTER ON clause, instead of failing with a subquery error. This fix is also included in the 2026.2.3 release.	Platform
N/A	When additional columns are added to a report with fixed column widths, those columns are displayed correctly when the report runs.	Platform
N/A	Running a High Availability repair now works correctly even if a SolarWinds service is unable to stop.	Platform
N/A	When a performance poll is manually triggered within a five-minute window, the SRM_StorageArrayStatistics_CS_cur tables are no longer set to NULL. (This issue relates to Everpure FlashArray, PowerMax, and Dell PowerStore arrays.)	SRM
02165798	Database maintenance no longer fails for HPE Allera 9000, HPE Primera, HPE Alletra MP10000, and HPE Greenlake for Block Storage devices because of a float.max value for the DataReduction and TotalReduction columns.	SRM
01972776	To prevent connection failures, attempting to connect to an HPE Modular Smart Array (MSA) (Redfish) using a web proxy is no longer possible.	SRM
N/A	The statistics poll can be triggered manually using the button on the Array Details page.	SRM
02137399	Disks that are polled with the ONTAP REST API are limited to disks that are local to the storage array, excluding disk from partner arrays in the MetroCluster. This prevents disks from being counted twice, which avoids unnecessary license consumption.	SRM
02123856	Database maintenance no longer fails on the SRM_StorageControllerStatistics_DetailToBaseline stored procedure with Arithmetic Overflow errors.	SRM
01855682	*The Virtualization Asset Summary resource no longer shows 0 for total physical CPU cores and total RAM when the environment includes Nutanix AHV hosts. AHV host resources are now counted correctly.	VMAN
02040954, 02038525, 02040126, 02053682, 02154496, 02153101, 02159785, 02171668, 02165627	*Having a large number of custom properties no longer causes the Manage Subnets & IP Addresses page from failing with a message similar to the following: An error has occurred. Unterminated string. Expected delimiter: '. Path '[39].DataType', line 1, position 4000. The underlying setting value is no longer truncated at 4000 characters. This fix is also included in the 2026.2.3 release.	IPAM
02083369, 02026720, 02010290, 02097812, 02108965	*Platform Connect now runs a preflight connectivity and TLS readiness check before UAMS registration, detecting common problems such as blocked outbound ports, DNS resolution failures, TLS trust-chain issues, and blocked ALPN/HTTP2 negotiation. If an issue is detected, it displays clear remediation guidance instead of a late, unclear failure.	Platform
02148084, 02168675, 02164466	*Node statistics collection no longer fails for an entire batch of nodes when a single node reports an unusually long ICMP response time. Response times that exceed the previous storage limit are now handled without causing bulk insert failures for other nodes.	Platform
02140909	*Adding a Proxmox VE node through an Additional Polling Engine no longer hangs at the Verify Credentials step.	VMAN
02142579, 02193615	*Nutanix datastore capacity values are now polled correctly for Nutanix v4-monitored clusters, and the Virtualization Storage summary page no longer shows invalid capacity figures.	VMAN
02135542	*The Report Builder Custom Table data source list no longer shows two data sources with the same name: VoIP Call Details. The active and historical VoIP call data sources are now distinguishable.	VNQM
02123590, 02147402, 02161418	*Users are now warned before uninstalling the UAMS plugin in a High Availability environment because doing so without first disabling HA can lead to failover issues.	Platform
02146292	HA pools are no longer automatically disabled as unlicensed on EOC installations that have the SolarWinds Observability EOC Node Attach Perpetual License.	EOC
02155686, 02119115, 02122152	*The Platform Connect wizard no longer fails to validate tokens and complete API calls when Internet access is through a proxy server. The wizard now respects the configured SolarWinds Platform proxy settings.	Platform
N/A	*The conversation length limit in the AI assistant chat has been increased so that conversations are no longer cut off after 10 turns.	Essentials
N/A	*The Node Details maintenance status information has been reformatted for readability when multiple planned maintenance jobs (mute, unmanage, and alert suppression) apply to a node at once.	Platform
02130240	*An audit log entry is now recorded when a user attempts to log in with an incorrect password through LDAP authentication.	Platform
02172391	*Firmware vulnerability matching now recognizes updated CPE string formats used by the National Institute of Standards and Technology (NIST). This change enables firmware vulnerabilities to be correctly matched to devices such as Windows 2022 servers.	Advanced
02200153	*The VNQM Add VoIP Gateway wizard no longer fails with an unhandled System.ArgumentNullException when the selected node has only SNMPv3 read-only credentials configured. A clear message is now shown instead.	VNQM
02195545	*The Configuration Wizard no longer reports failed alert imports for canned alerts that belong to unlicensed or inactive modules.	Platform
02178123	*The Cloud Monitoring AWS poller no longer accumulates large numbers of CLOSE_WAIT TCP connections over time. AWS client connections are now released promptly instead of only at the end of each polling cycle.	SAM
02170108	*The Platform Bridge UAMS plugin and other .NET-based UAMS plugins no longer randomly fail to start with a CryptographicException when parsing its client certificate.	Platform
02151147	*Adding an AWS account with EKS node groups created from a Launch Template no longer fails to parse polling data for the associated Kubernetes clusters.	SAM
02140822	*When an application monitor template with a process monitor component is added to an AIX node that is monitored with the SolarWinds Agent, polling no longer results in error messages being written to logs.	SAM
02182782	*Scheduled discovery configured on a Remote Collector no longer fails after the initial run. Subsequent scheduled discoveries now execute correctly.	Platform
02140909	*Proxmox VE and Nutanix nodes added through an additional polling engine while the main polling enging has no connectivity are now added and polled correctly.	VMAN
02103615, 02126355	*The Alerting Service no longer crashes when a background cleanup task encounters a recoverable SQL deadlock. Alert evaluation and cleanup continue running instead of stopping throughout the platform.	Platform
02176816	*An Intelligent Map widget using Map Name Format on a legacy Group Summary page no longer fails to load the map when a view limitation is also configured.	Platform
02188472	*When the Meraki Wireless API poller attempts to poll a wireless client that has no recentDeviceSerial value, the polling cycle no longer fails. Affected records are now skipped and logged instead.	NPM
02187246	*Database maintenance for SD-WAN tunnel statistics no longer fails with a divide-by-zero error for FortiManager and Versa tunnels. New tunnel statistics rows are now persisted with a non-zero weight value.	NPM
02203409	*Upgrades no longer fail because the deployment uses an older self-signed SolarWinds certificate signed with the SHA-1 algorithm. The Configuration Wizard now uses RabbitMQ instead of NATS as the message bus broker when the certificate has a SHA-1 signature.	Platform
02187246, 02143453, 02188472, 02103487	*Meraki API polling has been optimized to reduce API call volume. Switch interface discovery, status polling, and IP-address collection now use organization-wide bulk endpoints instead of per-switch calls. Device availability and other endpoint responses are cached with configurable TTLs. SD-WAN performance and tunnel-statistics polling is skipped for passive HA appliances. IP-address polling can now be disabled per node, and client-side rate-limit handling and reporting have been corrected.	NPM
02191321	*Adding or removing a group member whose entity (such as an IPSec tunnel) has a null status in SWIS no longer causes an Unexpected Website Error when saving group membership changes.	NPM
02115776	*A failed poll for one Proxmox VE host no longer incorrectly marks all other Proxmox clusters as Unknown/Unreachable.	VMAN
02195198	*The Exploit Prediction Scoring System (EPSS) score import no longer fails on servers configured with a Spanish (or other non-English) regional format. EPSS score values are now parsed and stored using an invariant culture regardless of system locale.	Advanced
02126290	*API poller status assigned to an external node is now correctly reflected in the node's status, and polling data is recorded as expected instead of the poller remaining in an unknown state.	SAM
02179634	*The Job Engine worker process no longer fails with an unhandled exception during WMI-based polling, improving background polling job stability.	Platform
02072344, 02082763, 02112987, 02113470, 02113423, 02114103, 02104559, 02102320, 02100879, 02115848, 02121125, 02142966, 02192207	*The JMX Bridge component used for Java/JMX application monitoring now ships with an updated version of Apache Log4j, resolving the publicly disclosed vulnerabilities previously flagged by security scanners against the older bundled library version. SolarWinds products were not affected by these vulnerabilities.	SAM
02196698	*Polling jobs no longer log XML serialization errors caused by invalid control characters in collected data, eliminating repeated error entries from the UDT job log.	UDT
02166379	*A time series widget that includes a PerfStack project can now be edited. Editing is no longer blocked because the Starting timespan field is required but no value can be selected.	Platform
02147435, 02157176	*The Edge Devices widget now displays the private management IP address for Fortinet FortiManager-polled Edge devices instead of the public WAN IP. Therefore, devices can be correctly matched with their corresponding SNMP-polled node.	NPM
02168198, 02170683, 02184168	*Saving changes to an assigned application monitor no longer fails with a server error when a threshold baseline has been set and template defaults are overridden.	SAM
01733844, 02121093, 02127266	*Email notifications now attach PDF files using the correct application/pdf MIME type instead of plain/text. Therefore, mail clients correctly recognize PDF attachments.	Platform
01943290, 01979718, 02174457	*When AppInsight for IIS is assigned to a node and the polling method changes, the AppInsight for IIS status is no longer reported as Unknown because the Log Parsing Monitor component failed. This fix is also included in the 2026.2.3 release.	SAM
01943737, 02026892	*When a Palo Alto firewall REST API key expires or becomes invalid, the Palo Alto poller no longer fails subsequent polls until polling is disabled and re-enabled on the affected devices. When a 403 response or credential expired message is received, the API key is automatically refreshed.	NPM

*This fix was added after the RC release.

Installation or upgrade

Known issues

SCM does not support gMSA credentials

Currently, SCM does not support gMSA credentials.

Resolution or workaround: None

HPE Alletra 9000/HPE Primera/HPE Alletra MP10000/HPE GreenLake for Block - SRM Maintenance Job Fails with "Arithmetic overflow error converting expression to data type real."

This issue has been fixed in 2026.4, but may still appear if you're upgrading from a previous version.

On some storage arrays, the REST API returns TotalReduction (and potentially DataReduction) as the real data type's float-max sentinel value (3.402823E+38) instead of NULL or a real reduction ratio (e.g. when the underlying ratio is undefined or divide-by-zero on the array side). When the timeseries maintenance job rolls this data up (Detail → Hourly → Daily), the aggregation arithmetic overflows the real range and the entire maintenance run for this table fails, blocking further rollups, resulting in the nightly/scheduled database maintenance job failing, and swdebugMaintenance.log showing:

Failed to execute 'Finalize Maintenance [SRM_StorageArrayCapacityStatistics_CS]' System.Data.SqlClient.SqlException (0x80131904): dbm_ExecuteMaintenance: dbm_AggregateData_part: dbm_AggregateTimeSerie_part: XACT_STATE=-1 Arithmetic overflow error converting expression to data type rea.l

You may also notice that historical/rollup data for SRM_StorageArrayCapacityStatistics stops updating (hourly/daily aggregation is stuck) once this failure starts occurring, since a failed maintenance run blocks subsequent aggregation for that table

Resolution or workaround:To fix this issue, contact support or perform a clean install of 2026.4.

End of life

Version	EoL announcement	EoE effective date	EoL effective date
2024.4	October 20, 2026: End-of-Life (EoL) announcement – Customers on SolarWinds Observability Self-Hosted version 2024.4 or earlier should begin transitioning to the latest version of SolarWinds Observability Self-Hosted.	November 20, 2026: End-of-Engineering (EoE) – Service releases, bug fixes, workarounds, and service packs for SolarWinds Observability Self-Hosted version 2024.4 or earlier will no longer actively be supported by SolarWinds.	November 20, 2027: End-of-Life (EoL) – SolarWinds will no longer provide technical support for SolarWinds Observability Self-Hosted version 2024.4.
2024.2	June 9, 2026: End-of-Life (EoL) announcement – Customers on SolarWinds Observability Self-Hosted version 2024.2 or earlier should begin transitioning to the latest version of SolarWinds Observability Self-Hosted.	July 9, 2026: End-of-Engineering (EoE) – Service releases, bug fixes, workarounds, and service packs for SolarWinds Observability Self-Hosted version 2024.2 or earlier will no longer actively be supported by SolarWinds.	July 9, 2027: End-of-Life (EoL) – SolarWinds will no longer provide technical support for SolarWinds Observability Self-Hosted version 2024.2.
2024.1	February 10, 2026: End-of-Life (EoL) announcement – Customers on SolarWinds Observability Self-Hosted version 2024.1 or earlier should begin transitioning to the latest version of SolarWinds Observability Self-Hosted.	March 12, 2026: End-of-Engineering (EoE) – Service releases, bug fixes, workarounds, and service packs for SolarWinds Observability Self-Hosted version 2024.1 or earlier will no longer actively be supported by SolarWinds.	March 12, 2027: End-of-Life (EoL) – SolarWinds will no longer provide technical support for SolarWinds Observability Self-Hosted version 2024.1.
2023.4	October 21, 2025: End-of-Life (EoL) announcement – Customers on SolarWinds Observability Self-Hosted version 2023.4 or earlier should begin transitioning to the latest version of SolarWinds Observability Self-Hosted.	November 20, 2025: End-of-Engineering (EoE) – Service releases, bug fixes, workarounds, and service packs for SolarWinds Observability Self-Hosted version 2023.4 or earlier will no longer actively be supported by SolarWinds.	November 20, 2026: End-of-Life (EoL) – SolarWinds will no longer provide technical support for SolarWinds Observability Self-Hosted version 2023.4.
2023.3	October 21, 2025: End-of-Life (EoL) announcement – Customers on SolarWinds Observability Self-Hosted version 2023.3 or earlier should begin transitioning to the latest version of SolarWinds Observability Self-Hosted.	November 20, 2025: End-of-Engineering (EoE) – Service releases, bug fixes, workarounds, and service packs for SolarWinds Observability Self-Hosted version 2023.3 or earlier will no longer actively be supported by SolarWinds.	November 20, 2026: End-of-Life (EoL) – SolarWinds will no longer provide technical support for SolarWinds Observability Self-Hosted version 2023.3.

See the End of Life Policy for information about SolarWinds product life cycle phases. To see EoL dates for earlier SolarWinds Observability Self-Hosted versions, see SolarWinds Observability Self-Hosted release history.

End of support

This version of SolarWinds Observability Self-Hosted no longer supports the following platforms and features.

Deprecation notice

The following platforms and features are still supported in the current release. However, they will be unsupported in a future release. Plan on upgrading deprecated platforms, and avoid using deprecated features.

Legal notices

© 2026 SolarWinds Worldwide, LLC. All rights reserved.

This document may not be reproduced by any means nor modified, decompiled, disassembled, published or distributed, in whole or in part, or translated to any electronic medium or other means without the prior written consent of SolarWinds. All right, title, and interest in and to the software, services, and documentation are and shall remain the exclusive property of SolarWinds, its affiliates, and/or its respective licensors.

SOLARWINDS DISCLAIMS ALL WARRANTIES, CONDITIONS, OR OTHER TERMS, EXPRESS OR IMPLIED, STATUTORY OR OTHERWISE, ON THE DOCUMENTATION, INCLUDING WITHOUT LIMITATION NONINFRINGEMENT, ACCURACY, COMPLETENESS, OR USEFULNESS OF ANY INFORMATION CONTAINED HEREIN. IN NO EVENT SHALL SOLARWINDS, ITS SUPPLIERS, NOR ITS LICENSORS BE LIABLE FOR ANY DAMAGES, WHETHER ARISING IN TORT, CONTRACT OR ANY OTHER LEGAL THEORY, EVEN IF SOLARWINDS HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

The SolarWinds, SolarWinds & Design, Orion, and THWACK trademarks are the exclusive property of SolarWinds Worldwide, LLC or its affiliates, are registered with the U.S. Patent and Trademark Office, and may be registered or pending registration in other countries. All other SolarWinds trademarks, service marks, and logos may be common law marks or are registered or pending registration. All other trademarks mentioned herein are used for identification purposes only and are trademarks of (and may be registered trademarks) of their respective companies.