Skip to main content

Frequently Asked Questions

Help

Short answers to common questions, with links to the full pages.

Account and plans​

What does the Free plan include?​

A 30-day trial with job scheduling, version history, the AI Builder and 1,000 AI credits a month. SAML single sign-on, log streaming, GitHub backup and organization groups are not included. Volume is capped at 2 connectors, 25,000 records per run and 50 scheduled runs per month, on a single environment and a Shared Cloud Agent.

How many AI credits does my plan include?​

Free includes 1,000 AI credits a month, Professional 5,000, Teams 20,000 and Business unlimited. One credit is 1,000 tokens of text sent to and returned by the AI model. See AI credits.

What happens when my AI credits run out?​

Milo, the Data Pipeline Builder, Map Fields and hosted form drafting stop answering until the credits reset, and AI Prompt nodes send rows to their Review port with the reason entitlement_denied. Runs keep going. Credits reset at the start of the next billing period, and upgrading adds credits at once.

What happens when my trial expires?​

Sign-in opens the Trial Ended page. Your data is kept, scheduled jobs are paused, and no new runs start. Choosing a paid plan restores access. See Trial ended.

Can I get my trial extended?​

The account Owner emails DataLug support at support@datalug.io.

Connections​

Why does Test Connection fail?​

The usual causes are a firewall between the runtime and the system, expired credentials, or environment-specific settings (a host or port that differs between DEV and Production). Test Connection needs a runtime mapped to the selected environment. The test lists each step it ran, and the first failed step shows where to look.

Can I move a connection to another account?​

Yes. Choose Export in the connection's row menu, then use Import on the Connections page of the other account. Secrets are encrypted in the export file. See Connections.

How are secrets stored?​

Passwords, API keys and OAuth tokens are encrypted and never returned to the browser. The connection form shows a masked value for a stored secret; leave the field as it is to keep the existing value.

Why doesn't my OAuth connection work?​

Three common causes:

  • Microsoft Entra requires the redirect URI type Web, not Single-page application (SPA).
  • The Google redirect URI must match the callback shown in the DataLug modal exactly, including case and trailing path.
  • A new Salesforce Connected App takes a few minutes to finish provisioning. Wait before testing.

Data Pipelines and nodes​

How do I combine several data pipelines?​

Add a Sub-workflow node to a parent data pipeline. Each Sub-workflow node runs another data pipeline as one step, in the order you wire them.

What does Quick Debug Run do?​

It reads 10 rows from the source, passes them through the flow, and previews the targets without writing. Use it to check transforms and output without touching production data.

How do I run a flow on a schedule?​

Open Jobs, select Add Job, pick the project, the data pipeline and the environment, choose a schedule mode, and save. Each scheduled run writes a Runtime Execution record. See Jobs.

Runtime and execution​

Where do I see what happened during a run?​

Monitor > Runtime Executions. Expand a run for its log, per-node metrics and, where the node records them, payload samples. See Runtime Executions.

Why is my run stuck?​

Open Environments, select the environment, and check the agent card on the Overview tab. An agent whose heartbeat is more than a few minutes old is offline, and the environment reads Runtime Unreachable. See Environments.

Can I stop a running job?​

Yes. Use Stop Run on the job's row on the Jobs page. A node that is mid-write may finish the rows in flight before the run stops.

Milo​

What can Milo do?​

Milo is the AI assistant in DataLug. It builds data pipelines on the canvas one question at a time, writes SQL queries, maps fields, explains failed runs and answers questions from the documentation. The Assistant only suggests; the Data Pipeline Builder changes the canvas as you answer its questions. See Milo in the app.

Does Milo use the connections I already have?​

Yes. The Source and Target questions list the connections this project can use. When a saved action on the chosen connection and object exists, the node uses it, so it opens configured.

Does Milo create connections or actions?​

New connection on a connection question opens the connection wizard. Milo does not create actions: a node without a saved action is finished in Configure Action after you save, which creates the action.

What does Milo see when I ask about a failed run?​

The run's status, project, environment, timings and execution type; the names and types of the nodes that failed; the error message; and the run log's diagnostic lines, which are the error and warning lines plus the per-node row counts.

It does not see the messages your Log nodes emit. Those are rendered from your message template and contain your row values, so they are withheld. The same applies to the message on a log event: Milo is told which node emitted an error, not what the record said.

Can I ask Milo questions on this docs site?​

Yes, where the Ask Milo button is shown. It uses none of your account's AI credits. See Milo on the docs site.

Connectors and integrations​

Do you support a system that is not in the catalog?​

Browse Connectors for the full catalog. A system with an HTTP API can usually be reached with the Generic API connector.

Why are my Google Drive or OneDrive files not showing up?​

The OAuth scope is the usual cause. Listing files needs read access; writing files needs write access, and the user must consent again after the scope changes. See the File hub setup guide.

Troubleshooting​

"Unexpected token '<' ... is not valid JSON"​

DataLug returned an error page where the app expected data. Retry. If it repeats on the same action, email DataLug support at support@datalug.io with the action and the time.

"Could not reach DataLug. Check your connection, then retry."​

The request did not reach DataLug at all. The usual causes are your network, a VPN dropping, or the computer waking from sleep. The panel showing the message has its own Retry; the rest of the app keeps working.

"Backend service is offline"​

A full-screen message, shown only when DataLug fails to answer twice in a row. Retry Now checks again immediately, and the screen clears when DataLug answers. Your work is not lost.

"Your session has ended. Sign in again."​

Your sign-in expired or was ended elsewhere. A session ends after 60 minutes without activity, or when you sign out in another tab.

When this happens while you are working, a Sign in to continue window opens over the page you are on. The page behind it is unavailable while the window is open. Sign in with any method your account allows: a one-time code, a passkey or single sign-on. The window closes and you stay on the same page. Retry the action that was refused. Unsaved canvas changes are kept.

Single sign-on leaves the page for your identity provider and brings you back to it afterwards. If you sign in as a different user, DataLug reloads from the start page instead.

Choose Sign out in the window to go to the sign-in page instead.

DataLug renews an active session in the background, including when you come back to a tab that was hidden. A session that has already ended is not renewed.

"That request took too long. Try again."​

DataLug stopped waiting for a reply. Nothing was saved. Retry. If it repeats on the same action, send that action and the time to DataLug support.

"Something broke on our end"​

An error inside the page itself. Retry reloads the page you were on, and Home returns you to the Dashboard. Reloading the browser also clears it.

I can't find the page I'm looking for​

Use the search box, browse the sidebar, or start from Getting started.