Fraznet Dashboard

Ad-hoc query runner

SQL Query

Label

Awaiting query…

Saved queries

No saved queries yet. Run a query and click ⊕ Save to save it.

Users

Username	Email	Role	Board	Last login

Boards

Board	Members	Widgets	Min recompute

API key (global)

Current key

New key

Notification channels — where alerts get delivered (Slack webhook or email)

Name	Type	Target	Enabled

Threshold rules — per-widget conditions that fire a channel on each ok→firing transition

Widget	Condition	Channel	State	Last fired

Recent firings — last 100 alert evaluations that crossed into firing

Time	Widget	Value	Threshold	Channel	Delivered

Query timeout — how long a MySQL query is allowed to run before being killed

The server kills any single SELECT that exceeds this many seconds. The default is set by the MYSQL_QUERY_TIMEOUT env var (currently —s). Raise it here if widgets are consistently timing out. Per-widget overrides can be set in the widget editor (global admins only).

Default query timeout (seconds)

Server caches — flush after editing a filter template, CPM rules, or any widget SQL referenced by these caches

The server caches two heavy reads. Filtered-snap caches widget results per filter-value combination so repeated filter clicks are instant. CPM eligible stores caches the latest-row-per-store scan of fraznetapp_cpm for the {{cpm_eligible_stores}}/{{cpm_latest}} tokens. Both refill on the next request after a flush.

HubSpot API connection — Private App token used by Python widgets

Paste a HubSpot Service Key token (or legacy Private App token). Widgets running Python get it as the HUBSPOT_TOKEN environment variable. Step-by-step setup ↗

Current token

Paste new token

HubSpot query library — reusable Python snippets users can drop into widgets

Name	Description	Updated

HubSpot dashboards (PNG snapshots) — snapshot URLs you can drop into Presenter mode

Label	Dashboard ID	Last snapshot	Pages	Dark mode	Kiosk URL

Other external dashboards (Genesys, Salesforce, generic) — snapshot URLs you can drop into Presenter mode

Label	Kind	Last snapshot	Pages	Dark mode	Kiosk URL

After adding a source you need to seed its login session once:
python bootstrap_session.py --source <source_id> on your local machine, then scp the resulting JSON to the server path shown.

Alignment

Left

Center

Drill-down filter — optional; clicking a bar / slice / cell applies its label as a value for this group filter

Annotations — optional vertical markers on the chart. One per line: x-value | label [| color]. Colors: red, amber, green, frazil, teal, slate, pink, orange.

Value scaling — how aggressively short numbers grow to fill the card

Query / Script

Rows — label | SQL, one per line (or script)

Max value SQL (for % bars)

Badges — label | SQL, one per line (or script)

Optional format per row: Label | SQL | format — e.g. integer, currency, currency_gbp, percent_raw.
Lines starting with --- become full-width section headers (e.g. --- This Week).

Layout Long values

Watched queries — label | SQL, one per line (or script)

Query / Script

Sort bars by

Query / Script

Chart style

Bar orientation

Vertical ▐

Horizontal ▬

Show values on bars/points

Useful in kiosk mode where hovering for tooltips isn't possible.

Show labels

Toggle the inline labels (region names on the map, slice values on charts, etc.). Tooltips and the colour fill stay regardless.

Query / Script

Max rows to display Highlight rules — optional; one rule per line: column OP value -> color. Operators: =, !=, >, >=, <, <=, contains. Colors: red, amber, green, frazil, teal, slate.

URL Method Headers (one per line, Key: Value) Request body (POST only, JSON) JSON path to value Display as

Path should resolve to a single value

Query / Script

Chart style

Pie

Donut

Show legend

Show

Hide

Target / max value

A fixed number or a single-value SQL query. The gauge fills proportionally from 0 to this target. The "Query / Script" above is the current value.

Query / Script

Two-column result: label, value. Order rows in the query — the first row is the top of the funnel.

Query / Script

One-column result: a numeric observation per row. The first column is used if more than one is returned.

Bin count

How many equal-width buckets to split [min, max] into. 2–100.

Layout Query / Script

Direction

Good if value reaches

Warn if value reaches

For "Higher is better": value ≥ Good → green, ≥ Warn → amber, else → red. For "Lower is better": ≤ Good → green, ≤ Warn → amber, else → red.

Status text

Optional text shown under the indicator: Good / Warn / Bad. Leave blank to show just the value.

Markdown text

Static text for context — no queries run. Supports headers, bold/italic, lists, links, and code spans.

Geocode by Query / Script

Lat/Lng: first two columns are latitude + longitude; third is label, fourth is the metric.

HubSpot Python script — runs server-side with the HUBSPOT_TOKEN env var available

Output type

Query timeout override — global admin only; 0 = use system default

Comparison query / Script

Runs alongside the main query on every refresh; delta shows current vs this value.

Export query / Script

If set, a download icon appears on the widget. Clicking it runs this query and downloads the results as a CSV file.

Groups with the same page render together in presenter mode. Skip numbers to leave gaps for future groups; gaps are ignored at runtime.

Overrides the widget title font-size for all widgets in this group. Useful for boards displayed on big TVs.

Group still shows on the dashboard but is skipped in the kiosk. Ungrouped widgets are always hidden from presenter mode.

Packs widgets top-to-bottom instead of left-to-right. Use this when you want narrow widgets to stack vertically next to a map/chart even if the grid has room to fit them side-by-side.

Copies the group, its widgets, and (optionally) its group filters. Widget alert rules and snapshot history are not copied — those are runtime state and may not belong on the target board.

Or custom: px (40–400)

Set password now Give the user their initial password directly. Email invitation to set their own Sends a one-time link so they pick their own password. Email is required.

Ad-hoc query runner

SQL Query

Label

Awaiting query…

Saved queries

No saved queries yet. Run a query and click ⊕ Save to save it.

Users

Username	Email	Role	Board	Last login

Boards

Board	Members	Widgets	Min recompute

API key (global)

Current key

New key

Notification channels — where alerts get delivered (Slack webhook or email)

Name	Type	Target	Enabled

Threshold rules — per-widget conditions that fire a channel on each ok→firing transition

Widget	Condition	Channel	State	Last fired

Recent firings — last 100 alert evaluations that crossed into firing

Time	Widget	Value	Threshold	Channel	Delivered

Query timeout — how long a MySQL query is allowed to run before being killed

Default query timeout (seconds)

Server caches — flush after editing a filter template, CPM rules, or any widget SQL referenced by these caches

HubSpot API connection — Private App token used by Python widgets

Paste a HubSpot Service Key token (or legacy Private App token). Widgets running Python get it as the HUBSPOT_TOKEN environment variable. Step-by-step setup ↗

Current token

Paste new token

HubSpot query library — reusable Python snippets users can drop into widgets

Name	Description	Updated

HubSpot dashboards (PNG snapshots) — snapshot URLs you can drop into Presenter mode

Label	Dashboard ID	Last snapshot	Pages	Dark mode	Kiosk URL

Other external dashboards (Genesys, Salesforce, generic) — snapshot URLs you can drop into Presenter mode

Label	Kind	Last snapshot	Pages	Dark mode	Kiosk URL

After adding a source you need to seed its login session once:
python bootstrap_session.py --source <source_id> on your local machine, then scp the resulting JSON to the server path shown.

Alignment

Left

Center

Drill-down filter — optional; clicking a bar / slice / cell applies its label as a value for this group filter

Annotations — optional vertical markers on the chart. One per line: x-value | label [| color]. Colors: red, amber, green, frazil, teal, slate, pink, orange.

Value scaling — how aggressively short numbers grow to fill the card

Query / Script

Rows — label | SQL, one per line (or script)

Max value SQL (for % bars)

Badges — label | SQL, one per line (or script)

Optional format per row: Label | SQL | format — e.g. integer, currency, currency_gbp, percent_raw.
Lines starting with --- become full-width section headers (e.g. --- This Week).

Layout Long values

Watched queries — label | SQL, one per line (or script)

Query / Script

Sort bars by

Query / Script

Chart style

Bar orientation

Vertical ▐

Horizontal ▬

Show values on bars/points

Useful in kiosk mode where hovering for tooltips isn't possible.

Show labels

Toggle the inline labels (region names on the map, slice values on charts, etc.). Tooltips and the colour fill stay regardless.

Query / Script

Max rows to display Highlight rules — optional; one rule per line: column OP value -> color. Operators: =, !=, >, >=, <, <=, contains. Colors: red, amber, green, frazil, teal, slate.

URL Method Headers (one per line, Key: Value) Request body (POST only, JSON) JSON path to value Display as

Path should resolve to a single value

Query / Script

Chart style

Pie

Donut

Show legend

Show

Hide

Target / max value

A fixed number or a single-value SQL query. The gauge fills proportionally from 0 to this target. The "Query / Script" above is the current value.

Query / Script

Two-column result: label, value. Order rows in the query — the first row is the top of the funnel.

Query / Script

One-column result: a numeric observation per row. The first column is used if more than one is returned.

Bin count

How many equal-width buckets to split [min, max] into. 2–100.

Layout Query / Script

Direction

Good if value reaches

Warn if value reaches

For "Higher is better": value ≥ Good → green, ≥ Warn → amber, else → red. For "Lower is better": ≤ Good → green, ≤ Warn → amber, else → red.

Status text

Optional text shown under the indicator: Good / Warn / Bad. Leave blank to show just the value.

Markdown text

Static text for context — no queries run. Supports headers, bold/italic, lists, links, and code spans.

Geocode by Query / Script

Lat/Lng: first two columns are latitude + longitude; third is label, fourth is the metric.

HubSpot Python script — runs server-side with the HUBSPOT_TOKEN env var available

Output type

Query timeout override — global admin only; 0 = use system default

Comparison query / Script

Runs alongside the main query on every refresh; delta shows current vs this value.

Export query / Script

If set, a download icon appears on the widget. Clicking it runs this query and downloads the results as a CSV file.

Groups with the same page render together in presenter mode. Skip numbers to leave gaps for future groups; gaps are ignored at runtime.

Overrides the widget title font-size for all widgets in this group. Useful for boards displayed on big TVs.

Group still shows on the dashboard but is skipped in the kiosk. Ungrouped widgets are always hidden from presenter mode.

Packs widgets top-to-bottom instead of left-to-right. Use this when you want narrow widgets to stack vertically next to a map/chart even if the grid has room to fit them side-by-side.

Copies the group, its widgets, and (optionally) its group filters. Widget alert rules and snapshot history are not copied — those are runtime state and may not belong on the target board.

Or custom: px (40–400)

Set password now Give the user their initial password directly. Email invitation to set their own Sends a one-time link so they pick their own password. Email is required.

Name Placeholder token — used in widget SQL as {{token}}. The inverse variant {{token_not}} is also available and substitutes as NOT (…) — handy for "everyone else" widgets. Filter type Filter expression template — {value} gets substituted with the selected value

Dropdown source SQL Value column — defaults to first column Label column — defaults to second column Include "All" option in dropdown Multi-select Picker becomes a checkbox list with Apply. Template uses {values} (plural) inside an IN (...) clause, e.g. c.chain_id IN ({values}).

Date range placeholders:
{start} / {end} — user input, empty when blank
{start_or_min} — start, or 1000-01-01 when blank (use for "before X" support)
{end_or_max} — end, or 9999-12-31 when blank (use for "after X" support)

Example template covering range, before, after, and unbounded:
c.date >= '{start_or_min}' AND c.date <= '{end_or_max}'

Require selection Removes the All option. Widgets using this placeholder sit in an empty "Pick a {Filter name}" state until a value is selected. Use this to gate expensive widgets so they don't auto-run on board load.

Object

Aggregation

Group by

Limit

Filters

Time window — optional

Property

Last N days

Python snippet — receives HUBSPOT_TOKEN as an env var; must print() JSON in the widget's shape

Source:

Kind:

URL:

Session file on server:

One-click bootstrap

Launches PowerShell on your laptop and runs the bootstrap automatically — no command paste needed. Requires the frazil-bootstrap:// URL handler to be installed once (see below).

🚀 Launch local bootstrap

First time? Download protocol installer · save the .reg file, double-click to install, confirm the Windows prompt. One-time setup per laptop.

Devices and browsers currently signed in to your account.

Add widget

Add group

Copy group

Add R script

Save query

Board settings

Add board

Add user

Edit user

Presentation mode

Add widget

Add group

Copy group

Add R script

Save query

Board settings

Add board

Add user

Edit user

Presentation mode

Filters

Set a new password

Add notification channel

Add alert rule

Add HubSpot query

Add HubSpot dashboard

Refresh login session

Add external source

My account