Release notes

This section describes the features, improvements, and bug fixes for the AWS Transform discovery tool, organized by release date.

Date	Version	Release notes
October 15, 2026	1.0.2810	New features Added SQL Server direct database collection. When you configure SQL Server database credentials, the discovery tool connects directly to SQL Server instances and collects database count, Always On availability group membership, cluster role, and read-only replica status. These fields appear in sql_server_data_full.csv. The tool falls back to Windows-level collection when no auto-connect credential can connect. SQL Server database credentials are configured on the Database access page. Added self-update. You can now update the discovery tool to a newer version in place without redeploying the appliance. Self-update preserves all collected data, discovery source configuration, and credentials. If the new version fails health checks after installation, the tool automatically rolls back to the previous version. Self-update is disabled by default and requires outbound internet access. Improvements You can now update the discovery tool in place with the Linux installer, keeping your collected data, credentials, and database encryption key. This is available for appliances on version 1.0.2325 or later; see Minimum version for each update method. Added VM skip lists for both VMware and Hyper-V. You can now exclude specific virtual machines from collection by editing a skip list file on the appliance. Both skip lists use the same file format, with glob-style pattern matching anchored to the full VM name. The skip lists also apply a set of built-in default patterns that exclude common infrastructure VMs (vCLS-*, NSX-Edge-*, NSX-Manager-*, and others). For the full list of default patterns, see Excluding VMs from collection. The discovery tool now skips non-routable addresses (link-local, loopback, unspecified, multicast, and localhost) when attempting OS-level collection. This prevents collection failures caused by unusable guest addresses and reduces unnecessary connection attempts. RFC 1918 private addresses are not affected. Each export ZIP now includes an export_summary.json file at its root. This file reports the overall status of the export (complete, partial, or failed), the export time range, and per-module counts of the servers, devices, and connections that were exported, along with details of any records that were skipped or could not be exported. If an export is partial or fails, the discovery tool directs you to this file for details. Oracle Database collection is now on by default. Previously you had to enable it explicitly. Kerberos authentication now builds the service principal name as WSMAN/<fqdn> by default, falling back to HTTP/<fqdn> if the server does not recognize the WSMAN SPN. This matches how Windows registers WinRM service principal names and reduces Kerberos authentication failures. Saving or reassigning a SQL Server or Oracle database credential now starts collection for the affected servers immediately, instead of waiting for the next scheduled round. The OVA and VHD appliance images now default to a 100 GB disk (up from 35 GB). The root volume is thin-provisioned on VMware, so the image still imports onto small datastores, and the file system grows to fill the disk on first boot. Bug fixes Fixed a SQL Server Reporting Services (SSRS) collection issue over WinRM that could cause SSRS components to be missed during collection. Fixed a timing issue where an on-demand ("Collect now") SQL Server collection running longer than one hour could have its in-progress status cleared early, letting that night's scheduled round start at the same time and corrupt or drop results. On-demand collection can now safely run its full two-hour budget. Fixed an issue where stopping SQL Server collection mid-round could mislabel a host's status as "Skipped (in backoff)" if that host's credential was deleted or rotated during the round. The host now keeps its prior status. The discovery tool now shows a Warning status when a SQL Server login is missing the VIEW ANY DEFINITION grant and Always On is enabled on the instance, instead of silently reporting a Success status with an availability group member incorrectly shown as standalone. Fixed the discovery tool's self-signed TLS certificate to cover the appliance's actual local network addresses, host name, and FQDN, instead of only localhost. The Discovered inventory page now records the last collection time for collection that you trigger manually, such as choosing Save and collect, importing a CSV file, or assigning a credential, not only for scheduled rounds. CSV import now also reports every credential name in the file that does not match a configured credential, instead of only the first one.
September 3, 2026	1.0.2325	New features Added disk monitoring and storage management. The discovery tool now tracks its own disk usage and warns you as space runs low. It pauses collection at 90% usage to avoid filling the disk. You can set how long the tool keeps data, for all modules or per module. You can also see usage by module and delete older data on the Settings page. Added unmapped_storage_devices.csv to the export. The discovery tool adds a device to this file when it cannot match the device to a mounted file system, or when the device reports no usable capacity. Improvements Improved Kerberos authentication for Windows servers. When the discovery tool knows a server's fully qualified domain name (FQDN), it now uses the FQDN first for Kerberos credentials. That order matches how Active Directory registers service principal names. You can still use IP addresses. The discovery tool maps the IP address to the FQDN with a reverse DNS lookup. Kerberos authentication now backs off after an unusable credential, such as an incorrect or expired password, instead of retrying every collection cycle. Threads that share one credential now make a single attempt between them. Both changes lower the risk of Active Directory account lockout. Added 95th percentile (_p95) columns to the OS metrics CSV exports. Renamed the disk throughput columns in the OS metrics CSV exports from _mbps to _mibps. The values do not change. They were already reported in mebibytes per second (MiBps). Improved the collection status for a virtual machine that reports no usable address. The discovery tool now reports why it cannot reach the machine, including the status of VMware Tools or Integration Services, instead of reporting only that no valid address was found. Added enablekrb5trace and disablekrb5trace aliases on the appliance to turn Kerberos trace logging on or off for troubleshooting. The discovery tool writes trace output to ~/.local/share/DiscoveryTool/logs/krb5_trace.log. Trace logging does not persist across a reboot, and a cron job rotates the log every 5 minutes to limit its size. Bug fixes Fixed Linux disk IOPS and throughput values. The discovery tool now samples a live interval instead of reading the average since the host last started, which underreported both values. Fixed exports that could fail on a large server inventory. The discovery tool now measures the memory available to its own service, estimates the size of network data more accurately, and runs one export at a time. Fixed the loss of collected network connection data when the discovery tool service restarts.
June 30, 2026	1.0.1646	New features Added Oracle Database discovery. The tool collects CDB and PDB topology, version, edition, installed options, feature usage, and component inventory. The discovery tool supports direct SQL connections (using SELECT_CATALOG_ROLE) and OS-level fallback detection through SSH or WinRM. Added support for multiple Kerberos credentials across different Active Directory domains. Each credential authenticates independently, enabling collection from Windows servers in multiple domains without authentication conflicts. Improvements Improved OS metrics collection for legacy hosts. Added fallback data collection for older operating systems (RHEL 5, SLES 11, Windows Server 2008 R2) where modern commands are unavailable. Storage, performance, and network interface data is now collected using alternative system interfaces. Bug fixes Fixed Kerberos authentication failures for servers discovered through Hyper-V. The discovery tool now correctly resolves the server hostname (FQDN) from Hyper-V guest metadata, which is required for Kerberos SPN matching.
May 14, 2026	1.0.1363	New features Added Linux installer for deploying the discovery tool on supported Linux distributions (Amazon Linux 2, Amazon Linux 2023, RHEL 8–9, Rocky Linux 8–9, AlmaLinux 9, Ubuntu 20.04–24.04, Debian 10–12, SLES 15 SP5). Added support for configuring multiple VMware vCenter servers with automatic VM deduplication across vCenters. Added SSH private key authentication as an alternative to username and password for Linux OS access. Supported key formats include RSA, ECDSA, Ed25519, OpenSSH, and PKCS#8. Added private address network collection to capture connections to and from RFC 1918 private IP addresses not in your discovered inventory. Added export options to select a custom date range and choose which data modules to include in the export. Improvements Improved scheduled collection reliability by staggering collector schedules. Improved OS metrics collection performance through SSH command batching. Credential changes now immediately retry previously failed servers. OS metrics collection now reports partial status when some commands return empty results. Improved error visibility for vCenter connection issues in the dashboard.
April 7, 2026	1.0.1074	New features Added support for Microsoft Hyper-V discovery, including auto-discovery of VMs from Hyper-V hosts with failover cluster deduplication. Added support for server import through CSV file upload. Added OS metrics collection with six independent modules: server performance, storage performance, server provisioning, storage provisioning, network interfaces, and running processes. Added Hyper-V deployment option using a VHD image file. Improved on-premises security hardening by removing EC2-specific services and blocking access to the instance metadata endpoint.
March 26, 2026	1.0.1014	Bug fixes and improvements Moved Kerberos domain capitalization validation from backend to frontend. Switched from Python-kerberos to Python-pykerberos and made file-based cache the default. Fixed case-insensitive bulk edit server ID validation. Fixed export ZIP filename (now discovery_tool_export.zip). Fixed service running check and added debug logs for system credentials. Added network configuration script to build artifact.
December 16, 2025	1.0.718	New features and improvements Added SMBIOS UUID support to prevent VM duplication in downstream services. Introduced Min/Max/Avg tracking and expanded storage metrics for improved resource monitoring. Added API validation improvements. Added help panel functionality with contextual guidance. Updated terminology from "collector" to "tool" throughout the interface.
October 31, 2025	1.0.0	Initial release Initial launch of the AWS Transform discovery tool. This release includes: Offline collection (no dependency on an AWS account or Application Discovery Service). VMware discovery and monitoring. SQL Server database discovery. Server-to-server communication collection through SSH, WMI, and SNMP.