Host Instance category
Nodinite alerts you the moment a BizTalk Host Instance stops, and lets the right people start it again from a Monitor View, without the BizTalk Administration Console.
- ✅ Every Host Instance on every server of the BizTalk group, as its own Resource
- ✅ Error when a Host Instance is stopped, Warning when it throttles
- ✅ Start, Stop, Restart, Details, and Group as Remote Actions, with role-based access and an audit trail
- ✅ A throttling chart per Host Instance, also as a Dashboard widget

Animated walkthrough: in a Monitor View filtered on the Category Host Instance, Stop on ReceiveHost32@NODINITE01 turns the row red (Error, "Host instance is NOT running"); Start brings it back to OK.
Understanding Host Instances Monitoring
Host Instances are the actual worker processes (BTSNTSvc.exe) that execute all BizTalk Server operations. Every Receive Location, Send Port, and Orchestration runs inside a Host Instance. If a Host Instance is stopped, the entire integration platform stops processing—no messages received, no orchestrations executed, no messages delivered. This is the single most critical component to monitor in BizTalk Server.
Why Host Instances monitoring is absolutely critical:
- Complete platform failure – A stopped Host Instance means zero processing for all artifacts assigned to that host. If your "ReceiveHost" instance stops, all Receive Locations stop accepting messages. If "ProcessingHost" stops, all orchestrations freeze mid-execution. If "SendHost" stops, zero messages are delivered.
- Silent catastrophic failures – BizTalk doesn't automatically alert when someone stops a Host Instance via BizTalk Admin Console or Services MMC. The entire platform can be down for hours before anyone notices (typically when partners start calling).
- Suspended instance accumulation – Stopped Host Instances cause messages to pile up in the MessageBox. When the instance restarts, thousands of dehydrated orchestrations and queued messages create rehydration storms that crash the server.
- SLA violations at scale – Unlike a single stopped Send Port (which affects one integration), a stopped Host Instance affects every integration assigned to that host. Orders, invoices, notifications, API calls—all fail simultaneously.
- Financial impact multiplied – One stopped Host Instance can halt hundreds of integrations. Every hour of downtime = lost orders, failed compliance filings, partner SLA penalties, and emergency escalation costs.
- Tracking data loss – Stopped Tracking Host means BizTalk stops writing tracking data to DTA database. Message bodies, orchestration debug traces, and BAM events are lost forever (cannot be recovered after restart).
- Cluster failover issues – In clustered environments, Host Instances must be configured correctly or automatic failover won't work. A misconfigured cluster means production goes down during hardware failure.
Common scenarios requiring Host Instance monitoring:
- Accidental stops – Administrator stops instance for maintenance and forgets to restart; entire host's workload ceases
- Service crashes – Memory leaks, unhandled exceptions, or resource exhaustion cause BTSNTSvc.exe to crash; without monitoring, instance stays down until manual intervention
- Throttling issues – Host Instance enters throttling state due to MessageBox database latency, high memory usage, or thread starvation; performance degrades to near-zero throughput
- Tracking host missing – No dedicated tracking host instance means tracking data isn't persisted; compliance and debugging impossible
- Host instance distribution gaps – Some BizTalk servers missing required host instances; workload imbalance causes some servers to be idle while others are overloaded
- Cluster configuration errors – Host marked for clustering but Windows cluster resource not configured; failover doesn't occur during outages
- Resource exhaustion – Host Instance consumes 100% CPU or memory but doesn't crash; processing slows to crawl without obvious errors
Host Instance types and their critical roles:
- Receive Hosts – Execute Receive Locations and adapters; if stopped, zero inbound messages accepted (orders, files, API calls rejected)
- Processing Hosts – Execute orchestrations; if stopped, all business logic workflows freeze mid-execution
- Send Hosts – Execute Send Ports and adapters; if stopped, zero outbound messages delivered (invoices, notifications, API responses never sent)
- Tracking Hosts – Write tracking data to DTA database; if stopped, all audit trails and debugging data lost
- Isolated Hosts – Execute in-process adapters (HTTP, SOAP); if stopped, web service endpoints return errors
Nodinite BizTalk Server Monitoring Agent monitors BizTalk host instances across all servers in your BizTalk group. The monitoring agent automatically detects and manages new or removed hosts and host instances, ensuring your environment stays up to date.
- Nodinite automatically lists all Host instances as Resources, where the name is the Host-name together with the configured server name, for example, 'SendHost32@SEBTPRD01'.
- Host instances belong to the BizTalk Group Name as their Application (not a specific BizTalk application).
- Host instances are further grouped by the following Category: Host Instance.
Note
Why do Host Instances use the BizTalk Group Name as Application Name?
Host Instances are system-wide resources that serve all BizTalk applications in the environment. They use the BizTalk Group Name (e.g., "BizTalk Group [Server\Instance]") as their Application Name to:
- Provide uniqueness when monitoring multiple BizTalk environments (Dev, Test, Prod)
- Distinguish system resources from application-specific resources (Send Ports, Orchestrations, etc.)
- Enable consistent filtering across all system-wide categories (Group, Host Instances, Health Check, Tracking)
See the Group page for a complete explanation of Application Name usage across BizTalk resource types.

A Monitor View filtered by the 'Host Instance' category.
What are the key features for Monitoring BizTalk Server Host instances?
- Cluster support – Verifies that your hosts are correctly set for a clustered environment
- Verify Tracking Host configuration – Alerts if a BizTalk server has no host instance with host tracking enabled
- Host instances exist for host – Alerts if not all host instances exist on each BizTalk Server node
- Running host instances – Alerts if a host instance is stopped (and not configured as disabled and/or part of a cluster configuration)
- Remote Actions – Supports the execution of remote actions
- State Evaluation – Monitors and evaluates the run-time state
- Metrics – Provides a throttling chart per Host Instance, also as a Widget for the Dashboard
What is evaluated for BizTalk host instances?
The table below describes the evaluated states for any host instance:
| State | Status | Description | Actions |
|---|---|---|---|
| 🚫 Unavailable | Resource not available | Evaluation of the 'BizTalk Host Instances' is not possible due to network or security-related problems | Review prerequisites |
| ❌ Error | Stopped and not disabled | A Stopped host instance renders as state 'Error', with the Log Text "Host instance is NOT running" (if not configured as disabled and/or part of a Windows fail-over cluster resource group) | Start |
| ❌ Error | Invalid tracking host configuration | A BizTalk server in the group has no host instance with host tracking enabled. The Resource is named Configuration error. No Host instance is configured for tracking @(server). New 6.4.2 Log Text: No host instance on server '(server)' has host tracking enabled. Nodinite recommends a tracking host instance on every BizTalk server, so tracking keeps working when a server is down and the tracking load is spread across the group. Enable 'Allow Host Tracking' on a host that has a host instance on this server, or add a host instance of the tracking host to this server | DetailsA BizTalk administrator must manually create dedicated tracking hosts to resolve this matter |
| ⚠️ Warning | Is throttling | Throttling is above configured thresholds. This behavior is built-in by BizTalk Server design, but may indicate exhausted resources—verify overall performance | Metrics chart: Throttling Metrics |
| ✅ OK | Enabled and running | Host instance is operational, and no other problems are detected | Stop |
| ✅ OK | Passive clustered host instance | The stopped host instance of a clustered host waits for a failover. New 6.4.2 Log Text: Passive clustered host instance in cluster group '(cluster)'; the active host instance runs on server '(server)'. Before 6.4.2, the Log Text said Host instance is NOT running next to the OK state | Details |
New 6.4.2 Clustered hosts also get these Log Texts; the states are unchanged:
- Only host instance in a cluster group – This host instance is clustered in '(cluster)', but it is the only host instance in this cluster group, so there is no failover. Add a host instance on another server to the cluster group
- Tracking on a clustered host – Tracking is enabled on a clustered host. A tracking host does not need clustering, because the tracking service runs on all its host instances at the same time. Use a separate, non-clustered host for tracking
Tip
You can reconfigure the evaluated state using the Expected State feature on every Resource within Nodinite.
What Remote Actions are available for BizTalk Server?
You can call actions on resources using the Nodinite Web Client tool or programmatically via the Web API. To execute remote actions, some Prerequisites must be satisfied. The user must log on to Nodinite using a Windows Credential and be assigned a Monitor Views with BizTalk Resources where Remote Actions are allowed. See Add or manage Monitor View for more details.

The Action menu of a BizTalk Host Instance.
One agent for every server in the group
One BizTalk Monitoring Agent covers the whole BizTalk Group; you don't need an agent per server.
- The agent reads the group's management database and lists every Host Instance of every server as its own Resource, named
Host@Server. - Start and Stop (and Restart, which is a Stop followed by a Start) control the Windows service of the Host Instance on the server named after the
@, through the Windows Service Control Manager. - The agent's service account needs the right to start and stop services remotely on every BizTalk server in the group: membership in the local Administrators group there, or a service ACL that allows it. Without it, the action fails with "Failed to start host instance ... there may be a permission issue". See the Prerequisites.
Metrics chart: What Metrics/Statistics are available for BizTalk Server?
Metrics chart: Host Instance Throttling
You can view individual and group (all instances combined) Throttling Metrics with Throttling chart under METRICS in the Action menu, or as a Widget on the Dashboard.

The throttling chart for a BizTalk Host Instance.
Filter
The BizTalk Monitoring Agent supports filtering unwanted resources. See the Configuration page for additional details.
Next Steps
Start
Stop
Restart
Details
Metrics chart: Throttling Metrics
Related Topics
Configuration of the agent
Group
Monitor Views