Local Cluster Troubleshooting v2

Use this checklist when a Workflow cannot select, save, Debug, publish, or run on a local Cluster.

An invalid Cluster is specified

Confirm the Cluster belongs to the current Site, permits the selected Environment, and is visible to your Cluster permissions. Refresh Local Clusters and the designer after creating or changing it.

Node or Connection is incompatible

The entire Workflow must run on one Cluster. Inspect every Node Method and selected Connection. A cloud-only contract or Connection assigned to another Cluster can prevent selection or execution.

Move incompatible work behind a separate Workflow and call it via Sub-Workflow, or use a compatible Node and Connection.

Cluster is offline

Open Local Clusters and confirm at least one host is online. On each affected Windows host:

Confirm Flowgear Local Runtime is installed, configured, and running.
Check outbound TCP 443, DNS, proxy, and certificate trust.
Confirm config.json contains the current Tenant host, Site key, Cluster key, and allowed Environment keys.
Review Windows Event Viewer and local runtime logs.

Certificate or authentication fails

Run the setup tool to create or locate the client certificate, upload its public .cer to the Cluster in Console, and confirm the displayed thumbprint matches. Do not upload a private PFX as the Cluster trust certificate.

After certificate replacement, restart the service and confirm a fresh host heartbeat.

Version or update mismatch

Compare host Versions on the Cluster detail page. Download the current installer and update hosts one at a time when high availability matters. Confirm the service is running after each in-place upgrade.

Workflow still will not run

Confirm publish, Environment status, Connection values, host capacity, trigger routing, and Sub-Workflow placement. Use Workflow logs to distinguish dispatch failure from a Step failure on the local host.