Skip to main content

Troubleshooting

Sign-in and session​

"Your session has expired. Please sign in again." — Sessions time out. Sign in again; you return to the page you were on.

"Workspace session expired" on a workflow screen, even after signing in again. — The app's workspace token is stale. Sign out from the profile menu, sign in, and re-select the workspace. If your workflows appear to be missing afterwards, confirm you are in the same workspace: the workspace switcher in the header shows the current one.

"Please verify your email before signing in." — Click the verification link. Use Resend verification email on the sign-in page if it has expired; check spam.

A social sign-in button says it is not configured. — That provider is not enabled for this deployment. Use email and password or Magic Link.

Workspace​

"Workspace not found" right after creating a workspace. — The workspace was created; the redirect raced the workspace list. Open Administration › Workspaces and select it. If it is still missing after a refresh, contact support.

"You don't have permission to do that." on the Dashboard. — One widget on the page needs a permission your role does not have; the rest of the page works. Ask a workspace admin for the role you need, or use Customize to hide the widget.

"Error loading tags" / "Error loading …" on a Workspace Modules screen. — A platform service behind that screen is unavailable. Retry after a minute. FluidGrids' own screens under Modules are unaffected. If it persists, report it through Support.

Building​

"Flow is invalid" when saving or running. — Every node must be connected and the graph must begin at a trigger. Look for a node with no incoming edge (other than the trigger) and connect or delete it. Sticky notes do not count.

"Workflow Limit Reached". — Your plan's workflow quota is used up. Delete unused workflows or upgrade under Billing. See Plans and Quotas.

A node cannot be configured / the panel is empty. — The node's manifest could not be loaded. Reload the page. If the node is a store node, open the Node Store and check that it is installed (it may need Update).

"No compatible node installed. Publish and install one from the Node Store first." — An AI Agent sub-node port (chat model, memory, tool) has nothing installed that fits it. Install the matching store node (for example OpenAI Chat Model).

"Unable to connect. Is the computer able to access the url?" — The builder could not reach the API. Check your network; if the app otherwise loads, reload the page. Repeated occurrences alongside other errors usually mean the session expired — see above.

"Credentials not configured" on a node. — Open the node and pick a connection under Connections, or add one under Credentials.

Runs​

The run is Failed. — Open it under Runs. The Timeline highlights the failed node and shows its error; the Logs tab has the full output. Common causes:

SymptomCauseFix
Authentication / 401 / 403 from a store nodeMissing, expired, or revoked connectionOpen the connection under Credentials, Test Connection, Refresh Token or reconnect, then Retry Node
HTTP Request node fails to connectWrong URL, service down, or the host blocks the requestCheck the URL and the service; add retries in the node's Settings
Expression errorAn ${output.…} reference to a field the upstream node did not produceOpen the upstream node's Output and correct the path
TimeoutThe run exceeded the execution time limitSplit into sub-workflows, reduce loop sizes, or raise the limit under Billing
Quota errorA fluidgrids.* quota is exhaustedSee Plans and Quotas

The run stays Pending. — All concurrency slots are in use. It starts when one frees.

The webhook did not start a run. — Webhooks run the Active version. Activate a version under Versions, then check the webhook URL on the trigger node (it is generated after the first save).

The schedule did not fire. — Same cause: only Active versions are scheduled. Also check the cron expression and the selected days of the week.

The run log stops with a limit notice. — The run exceeded its log limit and later lines were dropped. Reduce logging inside loops or split the workflow.

Still stuck​