Catch the failure before anyone looks.
Scout watches your Tableau Cloud or Server, detects the deviation, runs a repeatable triage playbook, and acts within your guardrails. So the 9am exec dashboard is already fixed - quietly - by the time anyone opens it.
Three modes. You stay in charge.
From off to autopilot - at your pace, not ours.
Most ops tools start at “automatic” and ask you to opt out. Scout starts at silent and earns its way to action - one approved play at a time.
Scout is installed but does nothing without your say-so. No signals, no actions, no surprises.
- Connected App auth verified
- Inventory synced from Tableau
- Audit log ready
Scout watches refreshes, queues, datasources. Files incidents with HeatScore. You decide if and when to act.
- RefreshWatchdog detector live
- Incidents triaged and prioritized
- Approval-required actions only
Scout acts on the lowest-risk plays automatically - within the guardrails you set per scope, tier, and owner.
- Auto-retry safe refreshes
- Notify owners on credential rot
- Escalate anything outside the guardrail
HeatScore
Severity is what broke. HeatScoreis what it's about to cost you.
Tableau ops teams drown in technically-severe alerts that are operationally trivial. HeatScore answers the only question that matters at 5:30am: what should I touch first?
It's consequence-weighted, not severity-weighted. Time, exposure, fan-out, and trust history - combined into a single number that means “fix this now” or “leave it for the morning standup.”
Heat cools overnight, heats up as scheduled view-times approach. A 5:30am refresh failure on a 9am exec dashboard is a 95. The same failure at 6pm is a 12.
An incident on a datasource feeding 40 workbooks heats faster than one feeding a sandbox. Scout walks the dependency graph automatically.
Scout knows which dashboards execs open, which finance touches at quarter-close, which marketing watches in the morning. Heat scales to who's about to look.
If Scout has lost trust on this asset before - false positive, missed regression, rolled-back action - that history dampens the score. Scout earns its weight back.
What's in the box
The four-part loop: Baseline → Detect → Triage → Act.
Watches every Tableau background job. Opens an incident on the 4th consecutive failure for any Tier-1 or Tier-2 datasource - no false-alarm spam.
Rule-based root-cause classifier: credentials, queue congestion, trust drift, downstream stalls. Each path maps to an action with explicit guardrails.
Every detection, every triage decision, every action - immutable. Compliance shows up tomorrow asking what you did at 3am? You hand them the log.
JWT Direct Trust via Tableau Connected App. Site-level for REST, tenant-level UAT for Platform Data API. No password storage, no OAuth chrome dance.
Autopilot only acts where you said it could. Restrict by site, by project, by datasource tier, by owner role. Anything outside escalates instead.
Notifications go where your team already lives. One incident, one thread, one acknowledgment - not a per-tool inbox per analyst.
Pricing
Free OSS core. Paid pilot for the full copilot.
The detection + audit foundation is open source - yours forever. The Autopilot, HeatScore, and integrations make Scout actually save you the 5:30am alarm.
- RefreshWatchdog + queue sentinel detectors
- Full triage playbook library
- HeatScore (consequence-weighted prioritization)
- Autopilot mode (with guardrails)
- Slack / Teams / PagerDuty / Jira
- Hosted deployment (we run it)
- Up to 5 sites + 50 owners during pilot
- Direct line to the founder
JWT auth, RefreshWatchdog, incident FSM, append-only audit, approval-required UI. Run it yourself. Releasing on GitHub day-of TC26.
Get notified when it ships →FAQ
Things admins ask before turning Scout on.
Sleep through the 5:30am refresh failure. Wake up to a fixed dashboard.
30-day paid pilot. We connect your Tableau instance the same day. If Scout doesn't earn its keep, cancel.