Skip to main content

Stratora v2.1.6 -- Node Stabilization & Upgrade Support

· 3 min read
Stratora Team

Stratora v2.1.6 introduces the node stabilization lifecycle, a new alert type for unresponsive agents, and full support for in-place MSI upgrades. This is the most significant release since the discovery engine overhaul in v2.1.0.

Node Stabilization Lifecycle​

Every monitoring platform faces the same problem after adding devices: the first few minutes produce a burst of false alerts. The node hasn't collected metrics yet, the health score drops to zero, and operators get paged for something that will resolve on its own in under a minute.

Stratora v2.1.6 solves this with a stabilization window. When a node is created -- whether through the discovery wizard, manual creation, or agent enrollment -- it enters a 5-minute stabilization period. During this window:

  • Alert evaluation is skipped for that node. No false "Node Unreachable" or "Agent Heartbeat Lost" alerts fire while Telegraf is still collecting its first data points.
  • The health score excludes stabilizing nodes. Adding 10 devices from a discovery scan no longer drops the infrastructure health score from 100% to 0% for five minutes.
  • The UI shows a distinct "Stabilizing" state with a blue pulsing indicator, so operators can see at a glance which nodes are still in their initial data collection phase.

When stabilization completes, the node transitions to normal evaluation and its first health status is computed from actual data rather than the absence of data. An audit log event records both the start and completion of each node's stabilization window.

Agent Not Reporting Alert​

A related problem: what happens when an operator adds a node expecting an agent to report in, but the agent is never installed? Previously, the node would sit in "Discovering" status indefinitely with no alert.

v2.1.6 adds the "Agent Not Reporting" alert. After the stabilization window expires, if an agent-type node has never checked in, this alert fires as a persistent warning. It tells the operator exactly what's wrong: the agent binary may not be installed, may be unable to reach the server, or may be misconfigured. The alert resolves automatically the moment the first heartbeat arrives.

In-Place Upgrade Support​

Previous Stratora versions required a clean install -- upgrading by running the new MSI over an existing installation could hang or fail. v2.1.6 makes in-place upgrades a supported path:

  • PostInstall.ps1 now detects whether it's running on a fresh install or an upgrade. On upgrade, it preserves existing database credentials, SSL certificates, and service registrations instead of regenerating them.
  • A PreInstall custom action stops the backend and NGINX services before the installer replaces binaries, preventing file-locking conflicts.
  • WiX ServiceControl elements ensure services are gracefully stopped during the standard MSI upgrade sequence on all future version transitions.

Upgrade Note​

Users upgrading from v2.1.5 may see a files-in-use dialog during installation. Click OK to continue -- no reboot is required. This dialog will not appear on future upgrades from v2.1.6 onward.

Technical Details​

  • Migration 150 adds stabilization_until and stabilization_completed_at columns to the nodes table with a partial index for efficient evaluator queries.
  • The alert evaluator's 4 separate grace-period blocks have been replaced with a single IsNodeStabilizing() function, making the stabilization check data-driven (database column) rather than computed from creation timestamp.
  • All 9 alert creation paths now explicitly check stabilization state, including the DNS resolution service which previously bypassed the evaluator entirely.

For installation and upgrade instructions, visit docs.stratora.io. The MSI is available on the v2.1.6 release page.