|Manuals

Instance and Pool Health

This article describes how to read the health of individual instances and of a model’s instance pool in Thinfinity® Cloud Manager. Health inspection is performed entirely through the Instances tab — each row is one instance, and the row’s columns and per-row actions are the primary inspection points.

The Instances tab is always scoped to a single model. To inspect another model’s pool, select that model in the left tree.

For the state values themselves, see Instance States and Lifecycle.

The Instances Tab

Select a model in the left tree to open the Instances tab for that model. Each row in the table is a single instance belonging to the selected model.

Default Columns

The table shows these columns by default:

Column What to Look For
(status icon) Green is healthy (Active). Red is Inactive. Yellow is a transition in progress. The alert overlay (⚠️) replaces the icon when the instance is in Error state. The forbidden overlay (🚫) appears on top when the instance is administratively Disabled at the permission layer.
Name The instance’s name. For VDI/DaaS pool instances, generated from the model’s Prefix, Digits, and Start Index in the General tab of the model editor; for Managed VMs, the name shown is the VM’s name as it exists in the provider.
IP Address The instance’s reachable IP, per the model’s Connection Mode. An empty or 0.0.0.0 value on a running instance suggests provisioning did not complete or networking was misconfigured.
Assigned Users Workspace users explicitly bound to this instance. Empty on a deployed instance in Allow Dynamic Assignment OFF mode means no user has been pre-pinned.
Active Users Workspace users currently connected. A stale Active Users value on an instance you expect to be idle may indicate an ungracefully ended session.
Actions Per-row state-aware actions (Start, Stop, Enable, Disable, View Log, Manage Users, etc.). See Manage Deployed Instances.

Click the Select Columns icon in the model header to add or remove columns. The current selection is saved per user, per model.

  • The Search field at the top of the table filters globally across Name, IP Address, Assigned Users, Active Users, and the model’s name / group / location.
  • Each column has a per-column filter accessed via its filter icon. The Status filter is a dropdown over the available status values; other filters are text matchers.
  • Use Reset Filters in the table toolbar to clear all per-column filters at once.

Recognizing Unhealthy Instances

Symptom Likely Cause Next Step
Status = Error (alert overlay) Provisioning, power, or destroy operation failed Click the alert icon to open View Log; see Provisioning Failures
Status = Active but no IP Provisioning incomplete, or cached state is stale Refresh the table (15-second auto-refresh; or use the Refresh button in the Cloud Manager top bar for an immediate poll)
Status stuck Provisioning for an unusually long time Terraform Apply is slow or hung on the provider side Open the model’s Deployments tab and check the in-progress deployment’s View Log; see Terraform Errors
Status = Active but users cannot connect Network, credentials, or Connection Mode mismatch See Connectivity Issues and Access Profile Errors
Forbidden overlay (🚫) appears on a healthy instance The instance was administratively Disabled If access should be restored, run the Enable action on the row

Pool Capacity Awareness

Pool capacity is assessed directly from the Instances table. For point-in-time pool size, count rows in the table and compare against the model’s Max Instances Count (configured in the General tab when Allow Pool Mode is enabled — see Provisioning and Pooling Options). For trend-based questions — are we hitting the ceiling regularly? — use the Thinfinity Workspace Analytics & Audit features for cross-resource historical analysis, or scroll through the Deployments tab and per-instance View Log entries for a per-resource history (see Instance and Deployment Logs).

When the pool is saturated (running instances at Max Instances Count ceiling, new users waiting), the available remedies are:

  1. Raise the ceiling. Edit the model and increase Max Instances Count in the General tab. Requires that your cloud provider’s regional quota covers the increase.
  2. Tune Staged Power On/Off. If enabled (Allow Pool Mode ON with Pool Mode = Breadth First / Depth First), reduce Occupancy Threshold to bring batches up sooner, or increase Additional Start Count to provision more instances per batch. See Staged Power On/Off.
  3. Reduce warm buffer. Lower Lazy Instances in the model’s Deploy tab if warm instances are tying up capacity that could serve users immediately.

For proactive alerting when pools approach capacity, monitor your cloud provider’s instance-count and quota metrics directly, or forward the per-instance lifecycle log to your monitoring stack. The Instances table is the inspection surface; alerting belongs to your external monitoring stack.

See Also