Data collection

Discovery tool collection schedule

After your initial discovery collection, the discovery tool continues to run on a staggered schedule to avoid resource contention:

The discovery tool also collects OS metrics through the following independent modules, each with its own staggered schedule:

You can independently start, stop, or trigger each OS metrics module by using Collect data now.

To manually run a collection, from the Actions menu choose:

These actions apply per module. You can control OS metrics modules individually.

OS data collection attempts

When a new server is discovered, the discovery tool attempts each configured credential for each IP address and the hostname. After the discovery tool finds a valid credential, it continues to use that credential unless you add a new credential.

After a collection failure, the discovery tool attempts to collect networking data for a server after 3 minutes, 30 minutes, 2 hours, and then 6 hours. After 4 failed attempts, the discovery tool continues to try all configured credentials once every 6 hours.

Discovered inventory

After you configure a discovery source, the Number of discovered servers value in the Discovery tool status frame begins to increment. The discovery status for the configured source changes to Enabled in the Collection module frame. The inventory page shows servers from all configured sources: VMware VMs, Hyper-V VMs, and imported servers. Each server shows its source and collection status per module.

Navigate to the Discovered inventory page to see the servers that the discovery tool has found. From this page, choose Download inventory to download a ZIP file (discovery_tool_export.zip) that contains up to 30 days of collected data, including MPA files for all configured sources, performance utilization data, database information, and server-to-server communication information.

You can download the ZIP file while the discovery tool continues to work, and obtain partial results. Upload this file to Migration assessment to obtain a business case for migration.

Export options

When exporting data, you can customize the export with the following options:

Date range

Select a start date and end date to export only data collected within that time period. Both dates are inclusive. The maximum date range is 30 days.

Module selection

Choose which data modules to include in the export. You can export all modules or select specific ones:

If you don't select any modules, the discovery tool exports all available data.

Data points collected

The discovery tool gathers comprehensive data across VMware, Hyper-V, OS metrics, database, and network components. The following sections detail the specific data points collected for each component.

VMware data collection

This table describes the VMware virtual machine information collected by the discovery tool:

Hyper-V data collection

This table describes the Hyper-V virtual machine information collected by the discovery tool:

Imported server data

Imported servers are not auto-discovered. They are imported through a CSV file. The discovery tool does not collect hypervisor-level data for imported servers. Instead, it collects database, network, and OS metrics data by using the OS credentials associated with each server during import.

Discovery tool's OS-related data

The discovery tool collects server inventory, performance, storage, network interface, and process data through SSH (Linux) and WinRM (Windows). The following tables describe the data points collected.

Server inventory (server_inventory.csv)

Combines server provisioning (hardware and OS configuration) with aggregated storage performance. Collected every 24 hours.

Server performance metrics (server_performance_metrics.csv)

CPU, memory, and network throughput utilization. Sampled every 10 minutes, aggregated over 30 days.

Storage performance (server_storage_performance.csv)

Per-volume disk I/O and space utilization. Sampled every 10 minutes, aggregated over 30 days.

Storage configuration (storage_config.csv)

Physical disk hardware details. Collected every 24 hours.

Network interfaces (network_interfaces.csv)

Network adapter configuration. Collected every 24 hours.

Running processes (process_metrics.csv)

Snapshot of running processes. Collected every hour, deduplicated over 30 days.

Network collection

The Network collection module helps you discover dependencies among servers in your on-premises data center. This network data accelerates your migration planning by providing visibility into how applications communicate across servers.

This module collects network data for servers from all configured sources, including VMware, Hyper-V, and imported servers. It uses WinRM to collect data from Windows servers and uses SSH, SNMPv2, and SNMPv3 to collect data from Linux servers.

Network data collection

The Network collection module captures TCP IPv4 connections in ESTABLISHED or TIME_WAIT state between servers in your discovered inventory. A connection appears in the output only when both the source and target IP addresses belong to servers that the discovery tool has discovered or that you have imported. Connections to or from IP addresses outside your inventory — such as external services, cloud endpoints, or servers not yet added to the discovery tool — are not included.

This design focuses the network data on server-to-server dependencies within your environment, which is the information needed for application dependency mapping and migration wave planning.

These data points are collected for each connection:

Private address network collection

By default, the Network collection module only captures connections where both endpoints are servers in your discovered inventory. You can enable private address collection to also capture connections to and from RFC 1918 private IP addresses (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16) that are not in your inventory.

To stop private address collection, open the Actions dropdown and choose Stop private address collection. Previously collected private address data is retained even after stopping.

Private address connections appear only in the full CSV export (inside the ZIP file), not in the MPA CSV files. If an IP address belongs to a server already in your inventory, it is always identified by its discovered server ID regardless of this setting.

This setting persists across restarts. You can start or stop private address collection at any time. Previously collected private address data is exported regardless of the current setting.

SQL Server collection

The SQL Server collection module gathers SQL Server information from Windows servers across all configured sources, including VMware, Hyper-V, and imported servers. The module uses WinRM to remotely connect to each Windows server and run PowerShell queries. It collects information about all installed SQL Server services (components) by using WMI namespaces, registry, and file properties.

A SQL Server component is a specific service or feature instance installed as part of a SQL Server deployment on a Windows server. The discovery tool collects Database Engine, Analysis Services, Reporting Services, and Integration Services.

SQL Server data collection

The SQL Server collection module gathers SQL Server component information. This table describes key data points collected:

Oracle Database collection

With Oracle Database collection, you can discover Oracle database instances across all configured sources, including VMware, Hyper-V, and imported servers. The module collects the following data:

Database-connected (SQL)

When you configure Oracle credentials, the discovery tool connects directly to the Oracle database using the read-only service account. This provides full SQL-level collection including PDB details, feature usage, and installed options. All CSV files contain complete data.

OS-level fallback

If you have not configured database credentials, or if the connection fails, the discovery tool uses SSH or WinRM to detect Oracle installations. It discovers Oracle homes, listeners, patches, version, and edition without database access. For OS-detected hosts, the exported CDB CSV contains instance name, host name, version, and edition. The PDB, features, options, and components CSV files contain no rows.

The Oracle collection produces the following CSV files in the export ZIP. Each table describes the data points collected per file.

CDB data (oracle_data_cdbs_full.csv)

One row per CDB instance. This table describes the CDB data points collected:

PDB data (oracle_data_pdbs_full.csv)

One row per PDB (pluggable database). This table describes the PDB data points collected:

Feature usage data (oracle_data_features_full.csv)

One row per feature usage entry from DBA_FEATURE_USAGE_STATISTICS. This table describes the feature usage data points collected:

Options data (oracle_data_options_full.csv)

One row per V$OPTION entry per CDB. This table describes the installed options data points collected:

Components data (oracle_data_components_full.csv)

One row per DBA_REGISTRY component. This table describes the component data points collected: