Highextract refreshes
3.9%refresh failure rate
45 of 1,166 extract refresh jobs failed
The failures are not spread across the schedule — they concentrate in six workbooks that fail repeatedly.
measured from: background_jobs · 90 days
Know exactly what your Tableau environment is doing.
Most Tableau health checks are a questionnaire. Ours is a measurement — taken from your own repository by a read-only script you run yourself. The numbers come back to us. Your data never does.
- Read-only
- No internet access
- You preview before you send
- No names, no credentials, no raw data
collector — bash
python vizbolt_collect.py --mode readonly --window 90d
vizbolt collector v1.4 · read-only · no internet access
connected workgroup@tableau-server:8060 as readonly
session pinned read-only — SET TRANSACTION READ ONLY
coverage report
background_jobs 90d 1,166 jobs → 96 hourly buckets
http_requests 30d 214,880 rows → TP50/TP95 load times
extracts now 1,412 objects → size + age bands
sessions 30d counts only → peak concurrency
object names replaced with codes — key stays on this machine
wrote ./vizbolt-summary.json (312 KB)
No connection left this machine. Preview the file,
then send it — only if you decide to.
Illustrative output.
Does your support plan include a technical review of your environment?
Whether anyone ever measures your Tableau environment usually depends on the support plan you bought — deep technical diagnostics are the kind of benefit worth checking your plan for, because many plans do not include one. If yours doesn't, that is not a reason to go without. It is a reason to get the same depth of review independently, on your schedule, from your own repository.
What an annual technical health review is — and how to get one →
Most Tableau assessments score your guesses. Ours measure your environment.
A questionnaire measures confidence. The repository measures reality. Every finding in a VizBolt review cites the query it came from and the window it was measured over.
The rule library behind those findings was built from 800+ findings across real enterprise Tableau reviews — so you also learn whether your numbers are normal.
Findings that cite their measurements
Every finding names its severity, its numbers, and the exact source and window they were measured from. These are illustrative examples from our rule library.
Highcapacity
19peak concurrent users
Peak concurrency exceeds provisioned VizQL capacity
Two VizQL processes served a measured peak of 19 concurrent users. By our own sizing convention — from practice, not vendor documentation — that layout is undersized for this load.
measured from: http_requests · 30 days
Mediumstorage
12.7 GBstale content
55 items have not been opened in over 90 days
Stale extracts sit in every backup and restore, extending maintenance windows for content nobody uses.
measured from: extracts · snapshot
Mediumbackgrounder queue
23 minTP95 queue wait
95% of scheduled jobs waited under 23 minutes — the worst waited 208
Delays cluster around 4 AM, where the schedule stacks refreshes into a single window.
measured from: background_jobs · 90 days
Illustrative findings — not from any single customer environment.
The findings become a scored, prioritized plan
Deployment Audit · report
74/100Watchlistdata coverage 86%
Inside the report
- Executive summary01
- Health Score & coverage02
- Findings & recommendations by pillar
- Performance & Reliability12 findings
- Infrastructure & Capacity7 findings
- Content Efficiency9 findings
- Data Flow & Scheduling8 findings
- Solution Design Quality6 findings
- Governance & Security5 findings
- Top 10 Actions: the 30-day plan09
- Data gaps10
- Appendices: full scorecard, Tableau workbook, method notes11
200+ measurements · six pillars · every finding paired with a recommendation
Sample highlights
6 workbooks cause most refresh failures
→ Fix credentials, re-point sources · wk 1
4 AM window overloaded
→ Move a third of the 6 AM batch · wk 2
55 items unopened in 90+ days
→ Archive after owner review · wk 4
31 workbooks have no identifiable owner
→ Assign owners before archive review · wk 3
Illustrative report fragment — not from any single customer environment.
Your data never leaves your network
The collector is read-only, has no internet access, and writes one aggregated file to your disk. You preview it; only what you approve crosses the boundary.
Collected — as aggregates
- Job outcomes, durations, and queue waits, bucketed by hour
- View load-time percentiles (TP50 / TP95) by view class
- Extract sizes and last-access age bands
- Session and peak-concurrency counts
Never collected
- User names and email addresses
- Workbook, project, and data source names (codes optional)
- Connection strings and credentials
- Query text of any kind
The full security model, including the collector's SQL, is on the how-it-works page.
What we assess
Six pillars, each read from the repository, the platform, or Admin Insights — not from a workshop.
Performance & Reliability
Extract refresh failures, TP95 load times, error patterns
Infrastructure & Capacity
Node topology, backgrounder sizing, peak concurrency against capacity
Content Efficiency
Stale content, duplicate extracts, storage growth
Data Flow & Scheduling
Live versus extract mix, refresh density, schedule overlap windows
Solution Design Quality
Custom SQL, calculated fields and LODs, image weight
Governance & Security
Permission complexity, dormant licenses, authentication posture
The full assessment covers far more — see the Deployment Audit.
Three steps, two weeks
- ≈ 20 minutes
You run the collector
Read-only, inside your network. It writes one summary file to your disk — you preview it and decide to send it.
- within 24 hours
We analyze
Your aggregates run against the rule library and the benchmark set. Analysis starts within a day of the file arriving.
- inside two weeks
You get the findings
Report, deck, workbook, and a prioritized action list — each finding citing the measurement that produced it.
Pricing
Published, fixed-scope, no discovery phase. Start free; pay when you want the full picture. Every paid engagement ends the same way: a scored, prioritized 30-day plan for your Tableau environment.
Health Snapshot
Free
no commitment
Run the collector, get your top five findings back within a business day.
- The collector, with instructions for your admin
- Your top five findings, with the numbers behind them
- A business day's turnaround
Deployment Audit
from $6,000
larger estates from $12,000
The full audit: every rule, every area, benchmarked against comparable estates.
- Findings report, executive deck, Tableau workbook
- Prioritized, owner-ready action list
- Delivered inside two weeks
Continuous Review
from $14,000/year
two reviews a year
Trends instead of snapshots: what changed, what regressed, and how you compare.
- Two full reviews a year
- Trend comparison between reviews, deltas quantified
- An anonymized peer view that sharpens as the program grows
Remediation is sold only after a review, at a day rate — never bundled, never assumed.
Start here
See exactly what your environment is doing
Start with the free Health Snapshot: run the collector, preview the file it writes, and get your top five findings within a business day.