scoutby TableauOps
Free assessment

The monitor for Tableau Cloud

Know what's on fire before your execs do.

Tableau Cloud won't tell you when an extract quietly stops refreshing, a subscription goes dark, or permissions drift. Scout watches your site around the clock, ranks what broke by what it's about to cost you, and calls it in to the right person with the diagnosis and the fix in plain English.

Get a free assessmentSee the live demo

read-only · 90-second assessment · real connected scan on request

Now

prod · 2m ago
2need you

Autopilot is handling 3. The other 39 targets are healthy.

Incidents · by heat
  • 94
    Q3_Sales_Exec
    Credentials expired · 38m · 212 viewers
  • 71
    Finance_Close_Daily
    Queue timeout · 1h 12m · 64 viewers
  • 42
    Web_Traffic_Hourly
    Transient 503 · 12m · 9 viewers
    RETRYING
Approve fixAsk Pilot

Three questions, three seconds

01

Is anything on fire?

One number, ranked by HeatScore. Consequence, not severity: the exec dashboard at 8:45am outranks the sandbox at 4pm.

02

Is Scout already handling it?

Every target is either healthy, being worked by Autopilot inside your guardrails, or waiting on a human. You can see which at a glance.

03

What do I need to do?

A plain-English diagnosis, the owner, and the literal next step. Approve the fix from Slack, your phone or the dashboard.

How it works

Not a dashboard you have to remember to check. A loop that runs against your site all day.

  1. 01
    Baseline

    Learns each target's normal: run time, cadence, who views it and when.

  2. 02
    Detect

    Watches every background job, publish rhythm and subscription for a deviation.

  3. 03
    Triage

    Classifies the root cause and scores it by what it's about to cost you.

  4. 04
    Call it in

    Routes the diagnosis and the fix to the owner, in plain English.

RefreshWatchdog

detect

Watches every Tableau background job. Opens an incident on the 4th consecutive failure for any Tier-1 or Tier-2 datasource, so there's no false-alarm spam.

Triage playbooks

triage

A root-cause classifier for credentials, queue congestion, trust drift and downstream stalls. Each path maps to a plain-English diagnosis and the literal next step in the Tableau UI.

Investigate

autopilot

When the one-liner isn't enough, open a forensic session on a single failure: concurrent failures, owner blast radius, schedule drift, upstream and downstream. Each session writes a Markdown case file. Read-only.

Append-only audit log

trust

Every detection, triage decision and call-in is immutable. When compliance asks what happened at 3am, you hand them the log.

Connected App auth

security

JWT Direct Trust via a Tableau Connected App. Site-level for REST, tenant-level for the Platform Data API. No password storage.

Guardrails by scope, tier and owner

autopilot

Autopilot only acts where you said it could: by site, project, datasource tier or owner role. Anything outside the guardrail goes to a human instead.

Routing where your team lives

call-in

Slack in every tier, managed Slack and email in Plus, and Teams, PagerDuty, Jira and ServiceNow in Autopilot. One incident, one thread, one acknowledgment.

HeatScore

Severity is what broke. HeatScore is what it's about to cost you.

Tableau ops teams drown in alerts that are technically severe and operationally trivial. HeatScore answers the only question that matters at 5:30am: which one do I call in first?

01
Time-to-view

Heat cools overnight and climbs as scheduled view-times approach. A 5:30am failure on a 9am exec dashboard is a 95. The same failure at 6pm is a 12.

02
Datasource fan-out

A datasource feeding 40 workbooks heats faster than one feeding a sandbox. Scout walks the dependency graph for you.

03
User and department exposure

Which dashboards execs open, which ones finance touches at quarter-close, which ones marketing watches in the morning. Heat scales with who's about to look.

04
Trust history

If Scout has cried wolf on an asset before, that history dampens the score. It earns its weight back.

When does this failure start to cost you?
958:00a · page someone now

Q3_Sales_Exec extract fails. 212 viewers, 9am exec review, 14 downstream workbooks.

drag across the bars or use ← → to move the failure

Modes

Scout watches. You decide how much it does about it.

A monitor that only pings you is noise. Scout makes the call on the deviations Tableau Cloud won't surface, and hands them to the right human. Autopilot is the paid action arm you switch on when you're ready to close the loop automatically.

Off
day 0 · autonomy 0

Scout is installed but watches silently. No Slack pings, no incidents opened, no surprises. The audit log stays empty.

  • Connected App auth verified
  • Inventory synced from Tableau
  • Audit log ready
Monitor
start here
day 1 · autonomy 1–2

Scout watches refreshes, queues and datasources. It opens incidents with a HeatScore, identifies the owner, drafts the diagnosis and calls it in to Slack with the fix in plain English. Eyes on, hands off.

  • RefreshWatchdog detector live
  • Plain-English diagnosis to Slack
  • Append-only audit log
  • Hosted dashboard and cross-site view (Plus)
Autopilot
when you trust it · autonomy 3–5

Close the loop. Autopilot runs the lowest-risk plays automatically, inside the guardrails you set per scope, tier and owner. Includes Investigate for one-failure-at-a-time deep dives. Same audit log.

  • Investigate for forensic deep dives
  • Auto-retry safe refreshes
  • Re-embed dropped credentials
  • Escalate anything outside the guardrail

Pricing

Self-hosted, hosted, or fully automated. One ladder.

Start self-hosted if Scout has to run in your environment. Add Plus for the hosted dashboard and managed alerts. Add Autopilot when you want failures fixed without a human in the loop. Every tier is a 30-day pilot.

paid · self-hosted
Scout
$99/ site / month
30-day paid pilot. Credited toward annual if you stay. Refunded if you walk.
  • CLI sentry: pull, hydrate, baselines, plain-English diagnoses
  • Runs entirely in your environment, on your infra with your keys
  • BYO Anthropic API key and Slack webhook
  • Read-only on Tableau
  • 1 site
  • Email support (48h response)
Book a setup call
We walk you through the install on the call. Most teams are running the same day.
paid · hostedrecommended
Plus
$249/ site / month
30-day paid pilot. Credited toward annual if you stay. Refunded if you walk.
  • Everything in Scout, hosted, with nothing to install
  • Hosted dashboard: connect your site at signup, live the same day
  • Up to 5 sites and a cross-site environment view
  • Managed Slack and email alerts (no webhook to wire)
  • 180-day retention
  • Email support (48h response)
Book a demo
Hands-on setup call on day 1. Autopilot unlocks as an upgrade during the pilot.
paid · fix
Autopilot
$799/ site / month
30-day paid pilot. Credited toward annual if you stay. Refunded if you walk.
  • Everything in Plus
  • Investigate: interactive forensic sessions on a single failure
  • Autopilot executor at autonomy 3–5 (confirm-each, auto-safe, full)
  • HeatScore (consequence-weighted prioritization)
  • PagerDuty, Teams, Jira and ServiceNow routing
  • Unlimited retention
  • Unlimited sites (volume pricing past 5)
  • Hosted deployment and a direct line to the founder
Starts inside your Plus pilot

No separate signup. Run the Plus pilot first, then switch on Autopilot when you're ready. We set the autonomy level with you.

The pilot fee is a deposit, not a bet.
Convert to annual and we credit the pilot toward the annual price. Walk away and we refund the full month. Email eric@tableauops.com. No tickets, no retention call, no questions.

FAQ

What admins ask before turning Scout on.

No - Scout is a scout. It watches, diagnoses, and calls it in to the right human (in plain English, in Slack) with the fix steps. There are two paid tiers on top: Plus is the hosted observability layer (managed dashboard, Slack alerts, environment view) and Autopilot adds the action arm that performs the safe plays plus Investigate, the interactive forensic REPL. That split is deliberate: it keeps Scout and Plus dead simple to security-review (read-only on Tableau, write-only to Slack) and lets you add execution when you trust it.

Stop being the first responder. Be the second pair of eyes.

Scout has the diagnosis waiting before standup. Self-hosted in your environment, or hosted with Plus. Add Autopilot and it closes the loop on the safe stuff while you sleep. If the pilot doesn't earn its keep, cancel for a full refund.

Start the Autopilot pilotOr start with Plus