Troubleshooting

Quick fixes for the problems people hit most: signing in, numbers that look wrong, uptime false alarms, missing alerts, billing, and agents.

Updated 4 min read Open LopeBase

Short answers to the problems people run into most, grouped by what you were trying to do. If yours is not here, see Fix a connection that needs reconnecting, or email us from the link at the bottom of this page.

Signing in and workspaces

Something else? Contact us.

I signed in but I'm in the wrong workspace

Switch workspaces from the menu next to your name at the bottom of the sidebar. If a workspace you expect is missing, you may have been invited under a different email than the one you signed in with.

My coding agent's sign-in link never opens an approval page

See Connect Claude and other agents over MCP: some terminals cannot open a browser on their own, so copy the printed link and open it yourself.

I signed out everywhere and my phone app still shows me signed in

Open the app again; it lands on the sign-in page instead of reloading. If it does not, close and reopen it once.

Numbers missing, wrong, or stale

Something else? Contact us.

A number says n/a

LopeBase has not read that number yet, or the connection behind it cannot report it at all, which is different from a connection that is failing. Open the connection to see which.

A page shows an old number after I know it changed

Pages read from numbers LopeBase already has, not live on every page load, so there is always a short lag. The "Updated" note under a page's numbers says how old they are, and a Refresh button asks for a fresh read right now.

A connection shows "Needs reconnecting"

See Fix a connection that needs reconnecting.

Two analytics tools show different pageview counts for the same property

Each one counts differently (cookies, bots, single-page apps). LopeBase shows each source separately rather than blending them, so pick one as your source of truth per property.

Uptime false alarms

Something else? Contact us.

Uptime said a service was down but it was fine

A single failed check is tried again before it counts as down, and a check against a vendor's status page or a shared endpoint can misfire. If your adapter's service checks use your own app's credentials, as the setup guide walks through, this is rare; mark a non-critical service as such so its failure shows as degraded instead of taking down the whole product.

I get an alert the moment a service blips

Site-down alerts fire right away on purpose. A connected tool that stops answering waits about an hour of continuous failure before it alerts, so a short blip does not page you; it still shows on Overview sooner than that.

A service I added shows as degraded, not down

That is the point of marking it non-critical: its failure is visible without reporting your whole product as down. Uptime still shows it under Needs attention.

Alerts not arriving

Something else? Contact us.

I turned on browser alerts but never get one

Check Settings, Alerts says the state for this browser. If notifications are blocked, allow them in your browser's site settings next to the address bar, then reload. Send yourself a test alert from the same page to confirm.

I'm not getting email alerts

Confirm the Email checkbox is on in Settings, Alerts, and that email is enabled on your deployment; the page says if it is not.

I only want to hear about serious problems

Set "How serious" to Needs action in Settings, Alerts. That drops early warnings like a bounce rate approaching a limit, and keeps things that are actually broken or blocking.

Billing

Something else? Contact us.

My product stopped updating

See Plans, trial, and billing: it is most likely paused because your plan does not have room for every product you connected.

I can't connect anything new

Check Billing for a past-due or suspended notice; new connections, operators, and dashboards pause until an unpaid invoice is settled.

An AI feature stopped working

You are most likely out of AI credits for the month. See AI reviews and credits for buying a pack or bringing your own key.

Agents and MCP

Something else? Contact us.

My coding agent's token stopped working

It expired, was revoked, or the admin who created it is no longer an admin of the workspace. Issue a fresh one from Connect Claude and other agents over MCP.

My agent says a tool needs a scope it does not have

Create a new agent token with that scope (setup, read, or manage) from Settings, Agent access, and point the agent at it.

npx lopebase can't find my coding agent

It looks for Claude Code, Codex, Gemini CLI, opencode, and Cursor Agent on your machine. If yours is not one of those, add the MCP server by hand; see Connect Claude and other agents over MCP.