Read your dashboard

The Dashboard is the first page after sign-in, and it answers three questions in one screen: is anything happening right now, did the last seven days work, and is anything broken. Everything on it is your agency's own data — no other tenant's numbers ever appear here.

The line under the title is the live header: today's date and time, then either "N calls in progress" (with a pulsing dot) or "No call in progress", then how many are in queue, then "last 7 days" — the window every KPI below uses. All times on this page are shown in Europe/Rome (the CET suffix next to the date).

Calls made from the in-browser agent tester are excluded from every number on this page. A test lead is not: it is a real call and it counts like any other.

The page is rendered fresh on every visit; there is no auto-refresh, so reload it to see the newest state.

Operational status

Right under the header sits the readiness block, the same nine checks as the Initial setup page. While something is missing you get an amber card with the warning "Some checks are not green yet: until they are, calls will not fire." and the full list. Once everything is green it collapses to a single line — nine dots and "Operational status: everything wired" — that stays there permanently and links back to the setup page.

That line is deliberate: if a wire breaks later (a number deactivated, a client deactivated, GoHighLevel disconnected) the full checklist comes back on its own instead of the problem going unnoticed. Each check is explained in The onboarding checklist explained.

Attention required

Below readiness, an amber "Attention required" card appears only when an operational alert has fired in the last 24 hours. Each line is one alert type, with a red dot when critical, a ×N counter when it happened more than once, and the time it was last seen. A repeated condition keeps the banner alive and its counter climbing; a durable record is written at most once every six hours per type, so the banner never floods.

There are exactly three alerts:

• "Call capacity near the limit (3+ hours at >=85% of max concurrency)" — for three or more consecutive hours your simultaneous calls sat at 85% or more of your maximum concurrency. Calls are not lost, but they queue instead of dialing. Raise the concurrency limit in Settings, or spread the load; see Understanding number saturation.
• "Some calls were permanently dropped from the queue and will NOT be retried" — an item was abandoned: either it failed to dispatch three times in a row (a provider or database failure that did not clear), or its agent is an inbound agent, which is never allowed to place a call. Nobody will call those leads unless a new trigger arrives. Check the agent's direction and run through Calls are not firing.
• "GoHighLevel connection broken - reconnect it from Connections" — GoHighLevel refused to renew your access token, so the CRM connection is dead: no contact reads, no tags, no notes, no calendar. The fix is always the same — open Connections and run the GoHighLevel connect flow again, as in Connect your CRM.

Every alert also leaves a durable alert.fired row in the audit log, recorded under the user system: see Understand the audit log.

The five KPI cards

Each card shows a number for the last 7 days, most of them with a small trend line:

• Outbound calls — calls your agents placed.
• Inbound calls — calls that came back in on your numbers.
• Conversion rate — booked appointments divided by outbound calls, with "N appointments booked" underneath. It reads — until there is at least one outbound call.
• Total cost — the sum of the per-call cost recorded for the window, in US dollars. A call whose cost was never reported by the provider counts as zero.
• Scheduled calls — this one is not a 7-day figure: it is how many calls are parked waiting for their moment right now (retries, agreed callbacks, leads held until their calling window opens), with "next at …" underneath, or "nothing waiting". It is the answer to "why is nobody being called at this instant?".

Live · now

The left card is the live feed: up to six calls that have not ended or failed yet, newest first. A pulsing dot and "in progress" mark a call actually running; a grey dot and "queued" mark one waiting. Each row shows the phone number with a direction arrow and the client it belongs to. The badge in the header repeats how many are active, and when there is nothing to show the card reads "No active or queued calls.".

Per client · 7 days

The right card is a leaderboard, not a directory: only clients that actually received a call in the window appear, most active first, up to eight rows with a "+N more clients with calls · see all" footer when there are more. Columns are Client, Calls, Appt. (appointments booked) and Conv. (conversion rate, highlighted in green from 10% up). The badge next to the title reads " of clients" — the second number is every client on file, including imported sub-accounts that have never been called.

The two empty states say different things. "Import your sub-accounts from GoHighLevel or create a client to get started." means you have no clients at all. "N clients on file, none has received a call yet. Check the readiness checklist above." means the clients exist but nothing has reached them — which is a readiness or workflow problem, not a dashboard one.

System health

The strip at the bottom shows five real signals, and the summary on the left reads "Operational" in green only when all five are up, "Attention" otherwise:

• Retell, Twilio, GoHighLevel — connected when valid credentials are saved for your agency, disconnected when they are not. This reflects what is stored in your vault, not whether the provider is having an outage right now.
• Postgres and Redis — active or offline from a real ping executed while the page renders. These are GodSetter's own database and queue: if either is offline, calls stop, and it is our problem, not your configuration.