SourceLace Docs
FAQOpen the app

All FAQ categories

Troubleshooting

Common messages and what to do about them. Each system's own page lists the messages specific to it.

Why does a source say "not connected" or "needs sign-in"?

Each person connects each source once, with their own login. Click Connect on the My connections page in the SourceLace app (or, if you use SourceLace from your own AI app, ask it to connect the source); a source that needs your sign-in has Connect right in its Status cell and is listed first; Needs you at the left and the Status list show only those. A Shared account source needs no sign-in page: one press of Connect, and only an admin can change it. Sign-ins can also end on the system's side, for example after a password change; then just connect again. A source marked Error failed its admin's last connection test: your admin fixes it.

Details: Your sources page and When something goes wrong.

Why is a system "not set up on SourceLace yet" or "not available yet"?

SourceLace's own app for that system is not ready yet, or that kind of source is not offered yet, or it is a Pilot connector (such as Marketing Cloud) that has not been switched on for your organization. Contact support@sourcelace.com.

Details: When something goes wrong.

It says a label is the label of two sources.

Two of your organization's connections were given the same label before each connection needed its own, so that label cannot pick one of them. For now, name the source by its connection ID, which the message lists. An admin fixes it for good by renaming one of them: Connections shows a notice for each shared label, with Edit next to each connection.

Details: Several connections of one system.

I am told I am not allowed to sign in. What now?

Your organization's SourceLace admin needs to invite you or add your email domain. If your organization requires single sign-on, choose Continue with SSO / work email and sign in through your identity provider.

Details: Your first five minutes and Single sign-on.

It says my groups do not give me access.

Access by group is on, and none of your groups allows that source or object. Ask a SourceLace admin to add you to the right group.

Details: Access by group.

How do I check that my connections and the AI assistant are working?

Open My connections. The bar at the top shows whether the AI assistant is ready, and each line shows whether you are connected and how your calls went in the last 24 hours. Needs you, in the list at the left, shows what needs your sign-in or has a problem, and the Status list above the table narrows it to Needs sign-in or Problems. The old Connection status page is part of this page. If something is not working, the page links to the Help centre.

Details: Status, usage and invoices.

How do I find out which of our connections or agents is failing?

Admins open Health (under Administration, Trust and health). It lists, problems first, each connection that failed lookups in the last 24 hours, agents whose runs failed this week, whether the daily audit check passed and whether each streaming destination is delivering. Open Connections from a failing connection to run Test connection. Everyone else sees their own sources, and how each is working for them, on My connections.

Details: Health.

It says a result has expired.

Results are held for 30 minutes at most, or less if your admin chose less, then deleted. Ask the question again and SourceLace reads the system again.

Details: Data retention.

It says a change has expired or was already applied.

A change preview lasts 10 minutes at most and applies once. Ask for a new preview.

Details: Changing records.

The assistant in the app stopped answering. Why?

Chat says why. Common reasons: your organization reached its maximum AI use for the month (an admin can raise it, or add your own AI key), your own AI key was refused or ran out of credit, or the AI is busy (try again after the seconds it says). Your own AI apps keep working either way.

Details: AI settings and AI use: credits and questions.

One chat says the AI model "could not continue this chat", but new chats work.

The AI model turned down that one conversation, so trying again in it gives the same answer. Start a new chat: your other chats and new ones are not affected. If the same chat is too long for the model, the message says so instead; open the gauge at the top of the chat and press Compact all (or New chat from summary), or start a new chat. A chat that is nearly full normally compacts itself before this happens, unless your admin or you switched that off. If it keeps happening, send support the reference in the message (it starts with req_). The reference lets support find what happened; it holds nothing from your chat. When you reopen such a chat, the reply that failed shows "(No answer: ...)" with the reason.

Details: AI settings.

It says a system or the AI is busy: "try again in N seconds".

SourceLace shares each system's request budget and the AI's fairly between everyone in your organization, so that no one person or agent floods a system. That budget was full for a moment. Nothing went wrong and nothing was lost: try again after the seconds the message gives (your AI app can do this by itself). AI agents start again later on their own. If it happens often, an admin can see on which days on the Usage page.

Details: When a system is busy.

A system could not be reached or did not answer in time.

The system was down or slow, or its firewall blocked SourceLace. Try again shortly. For systems you run yourself, check that the firewall lets in SourceLace's outbound IP addresses.

Details: What your network needs.

Single sign-on sign-in fails.

The message says why, for example a client secret that expired, an email domain not ticked for single sign-on, or an address that already belongs to another organization. The manual lists each message and its fix.

Details: When something goes wrong.

An AI agent run failed.

You see a notice in the app, with the reason and a link to the run. A run stops when it reaches one of its limits, when the owner lost access to a source, or when an admin stopped agents. Fix the cause, then run it again or wait for its next run.

Details: When a run fails.

Data 360 sign-in fails with OAUTH_APPROVAL_ERROR_GENERIC, but Salesforce works.

Data 360 asks Salesforce for two more scopes than Salesforce does. Your Salesforce admin checks that SourceLace's app in the org lists the Data 360 scopes, and installs the current version from support if not. If the app only lets approved people in, the admin also checks the person has the SourceLace permission set. Then connect again.

Details: When something goes wrong with Data 360.

Salesforce sign-in fails with OAUTH_APPROVAL_ERROR_GENERIC or "user is not admin approved".

Your Salesforce org only lets approved people use SourceLace's app, and you are not one of them yet. Your Salesforce admin gives you the SourceLace permission set. Then connect again.

Details: When something goes wrong.

A database "could not be reached".

Most often the database's firewall does not let SourceLace's IP addresses in, or the host or port is wrong. Your database administrator checks the firewall rule (or the cloud's security group, firewall or IP access list) and that the database has a public address. "Could not be reached securely" means the certificate could not be checked: your SourceLace admin adds the CA certificate to the source.

Details: When something goes wrong.

A database does not accept my username and password.

Check them with your database administrator: each person uses their own database login. If it worked before, the password may have expired or the account may be locked; the database's own log says which. On MongoDB Atlas, sign in with a database user from Database Access, not the email you use for the Atlas website.

Details: Codes in the database's own log.

SAP says it refused my new sign-in, and connecting again does not help.

SAP refused even a sign-in it had just given you, so its setup does not let you in yet. Through SAP BTP, your BTP and SAP Basis teams check principal propagation (the destination, Cloud Connector, and STRUST and CERTRULE in ECC). In SAP Datasphere, your Datasphere admin checks that you have a Datasphere user and that SourceLace's client is Interactive Usage.

Details: SAP ECC: When something goes wrong.

Snowflake says it has no warehouse to run my question on.

Your Snowflake user has no default warehouse, and the source names none. Your SourceLace admin sets the source's Warehouse, and your Snowflake admin lets your role use it; or your Snowflake admin gives your user a default warehouse.

Details: When something goes wrong with Snowflake.

All FAQ categories