Every number on a Qvasa dashboard can be a sensor. Point a threshold at a widget and Qvasa watches it on a schedule; when it breaches, Qvasa sends a message — to Slack, to an email list, to a webhook of your own design, to SMS or PagerDuty — with the numbers, the breaching segments, links to the tickets behind them, and a rendered image of the dashboard attached. Separately, any dashboard can export itself as an image on a cadence and deliver it to the same kinds of destinations, so a morning snapshot lands in a channel before anyone opens the app.

Both are set up from a bar at the top of the dashboard itself — no separate admin screen, no scripting.

Open Alerts & Exports in the sidebar. The page lists every alert dashboard on the account with its conditions, current status (Alerting or OK) and last breach; click Create New Workflow Alert Condition (Dashboard) to start a new one.

Build it like any other dashboard: add rows, add widgets, set each widget's filters and date window. The window matters — a widget watching "the last hour" is what makes an alert about right now rather than about the month. Then click Edit in the top bar and click the bell icon on a widget to give it a threshold.

Once a widget has a condition, it says so on the dashboard: you can read every threshold on the board at a glance, and the command center bar across the top shows the alert's status, its schedules, its notification channels and its alert message.

You can put up to 5 conditions on one alert dashboard. They are ANDed: on each run, Qvasa checks them in order and notifies only when every condition is met, so a board with two conditions is "alert me when both of these are true at once".

Conditions attach to the numeric widget types in the library — hundreds of metrics across tickets, agents, messaging, queues and your own custom fields. There are three shapes, and each evaluates differently.

Single-value quick stats. One number compared against your threshold — tickets created in the last hour, agents online, median first reply time, backlog, anything the widget library counts as a single figure.

Formula quick stats. The five calculated quick stats — percentage, difference, sum, product and quotient — build a number out of two other widgets on the dashboard, then behave exactly like a plain quick stat. This is how you alert on a ratio nothing in the library measures directly, such as solves ÷ new tickets as a solve rate. (A quotient that divides by zero evaluates as 0.)

Grouped metrics — donut and horizontal bar charts. A grouped widget is many numbers at once: new tickets by priority, backlog by group, handle time by channel. One condition covers every segment. Qvasa compares your threshold against the single highest segment, so "above 40" means "alert me if any group goes over 40" — you don't set up one alert per group, and you don't miss a new group that appears later. When several segments are over the line, the alert names all of them (the top 5 by value) with their numbers, so the message tells you which queues are on fire, not just that something is.

On a grouped widget, a moving-average condition is evaluated per segment: each group is measured against its own baseline, and the alert fires if any group clears both checks — so a small queue doubling is caught even while a big queue looks normal. Segments with no history to average against can never breach.

In the command center bar, click Channels under Notifications and pick up to 4 destinations:

Each selected channel has an Also notify on resolve toggle. Turn it on for the channels that should hear the all-clear — typically the Slack channel gets both the breach and the recovery, while the webhook and the pager only get the breach.

The optional Capacity changes section goes one step further: an alert can apply a capacity rule when it breaches and restore the previous one when it resolves, so a queue spike can widen agent capacity automatically.

Click Alert Message to open the editor. A live preview at the top shows exactly what the breach and the recovery messages will say.

Below that, Alert Body Content controls which generated pieces ride along. Turn off what your team doesn't read; keep what it acts on.

When the alerting widget inspects tickets, Inspected Ticket Details adds the specifics behind the number: the Zendesk ticket ID list, a clickable agent link per ticket, the current assignee's email, an @-mention of each assignee in Slack via their saved Slack user ID (Slack only — email and SMS never see mention markup), and a custom JSON payload per ticket shaped by a ticket payload definition.

One rule: the payload can't be emptied. Either keep the breach description or write a preamble of your own.

A webhook destination sends whatever JSON shape your endpoint expects. Create one under Outbound Webhooks: an HTTPS destination URL, an authentication strategy (none, HMAC-SHA256 signing, bearer token, basic auth or a custom API key header — all credentials stored encrypted), and the payload itself.

The payload is a JSON template. Reference any value the alert produces with {{ value }}, and search the value list to insert one at your cursor: the workflow and condition names, the breach description, the trigger direction, the threshold and the observed value, the inspected Zendesk ticket ids and their per-ticket payloads, the dashboard URL, and any capacity changes the alert made. A token that is the whole string is replaced with the real typed value, so numbers stay numbers and the ticket payloads block stays a nested object. Start from the default envelope or restructure it entirely — only four top-level keys are fixed (event_type, event_id, occurred_at, account_identifier) so receivers can rely on a stable envelope. A preview renders the exact body that will be POSTed, optionally filled with real ticket data.

Deliveries are attempted once, and Qvasa tracks successes, failures and the last error. A webhook auto-disables after the consecutive-failure count you set (or immediately on a 410 Gone), and the owner is emailed when it does.

Exports are the other half, and they need no conditions at all: on a cadence, Qvasa renders the dashboard as an image and delivers it. Open any dashboard and click the calendar icon next to its title to open the export bar.

An export won't run until it has both a schedule and a destination — the status section says which one is missing. Each dashboard has one scheduled export.

Alerts and exports share one scheduling engine: cron expressions in UTC, up to 3 schedules each, added with + Add Schedule in either bar. Alerts can evaluate as often as every minute; the schedule doubles as quiet hours, since an alert that isn't scheduled to run simply doesn't. A common pattern is two schedules on one alert: every 30 minutes through business hours, hourly overnight. For the cron field reference, presets and timezone recipes, see Scheduling alerts and dashboard exports with cron schedules.

On the alert dashboard, the status section reads Active with a next run time, each widget shows its threshold, and the Notifications section lists your channels. On an exporting dashboard, Export Status reads Exporting with a next export time and your destinations listed. From then on the dashboard works whether or not anyone is looking at it: it watches the numbers, and it sends them where they need to go.