> ## Documentation Index > Fetch the complete documentation index at: https://docs.vortexiq.ai/llms.txt > Use this file to discover all available pages before exploring further. # Workflow Error Spike, HubSpot > Surge in failed workflow actions, usually a deleted property or unpublished email. Audit HS02 cross-fires. How to read it, why it matters, and how to act ... **Metrics type:** [Key Metrics](/nerve-centre/overview#metrics-types-explained) • **Category:** [Email Marketing](/nerve-centre/connectors#connectors-by-type) > Surge in failed workflow actions, usually a deleted property or unpublished email. Audit HS02 cross-fires. ## At a glance > Real-time anomaly detector for workflow errors. Fires when 24-hour error volume exceeds 2 standard deviations from the trailing 7-day baseline. Most-common causes: a deleted property, an unpublished email referenced by a workflow, a removed list trigger, or a send-from address that lost verification. | | | | -------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **What it counts** | Total workflow-error events in the last 24 hours, scored against the 7-day baseline error-event distribution. Each row in the alert table is one workflow plus its top-3 error categories. | | **HubSpot Hub scope** | All Hubs that allow workflows (Marketing Hub Pro/Enterprise, Sales Hub Pro/Enterprise, Service Hub Pro/Enterprise). | | **Threshold logic** | `(today_error_count - 7d_mean) / 7d_stdev >= 2.0`. Two-sigma positive move. Configurable per portal. | | **Error categories tracked** | MISSING\_PROPERTY, MISSING\_LIST, INVALID\_EMAIL\_ADDRESS, SEND\_FROM\_NOT\_VERIFIED, RATE\_LIMIT\_EXCEEDED, WORKFLOW\_VERSION\_MISMATCH, UNPUBLISHED\_EMAIL, MISSING\_TEMPLATE. The full list is documented in HubSpot's workflow-history reference. | | **Distinction from `Workflows With Errors`** | `Workflows With Errors` is the 7-day rolling list of any errored workflow. This card is the spike alert: a sudden burst above baseline. The two together give you "what is broken" plus "what just broke now". | | **Persistent vs flash errors** | Persistent failures contribute to the 7-day baseline, so they are normal-error-rate. Flash errors (a property deleted at 09:00 today) cause the spike. The card is designed to surface change events. | | **Multi-portal** | One card per portal. | | **Currency** | n/a (count). | | **Time window** | `24H` rolling. Alert fires within 5 minutes of breach. | | **Alert trigger** | `errors >2σ vs 7D baseline`. Audit HS02 cross-fires when the spike has a clear root-cause (single property delete, one workflow accounts for >70% of new errors). | | **Roles** | marketing, engineering | ## Calculation Calculated automatically from your HubSpot data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading. ## Worked example The same B2B SaaS portal with 38 active workflows. The 7-day baseline: | Statistic | Value | | ---------------------------- | ------------------- | | 7-day mean daily error count | 47 errors | | 7-day std deviation | 18 | | 2σ positive threshold | 47 + 36 = 83 errors | Reading on 14 Apr 26 morning: | Trigger event | Workflow | Error category | Count (24h) | | ------------------------------------------- | ----------------------- | ------------------ | ------------------------------------ | | Property `trial_end_date` deleted yesterday | Trial Expiry Nurture | MISSING\_PROPERTY | 142 | | Email "Welcome step 2" unpublished | Welcome Series (5-step) | UNPUBLISHED\_EMAIL | 58 | | Other workflows (steady-state) | (various) | (mixed) | 22 | | **24h total** | | | **222 errors (9.7σ above baseline)** | Five things this picture reveals: 1. **9.7σ is far above the 2σ threshold; the spike is unambiguous.** The audit log shows the property-delete event happened at 16:42 yesterday, just before close-of-business; the next 16 hours of trial-expiry enrolments all failed at the delay step that referenced the missing property. Audit HS02 fired at 17:38 yesterday once the delay-step backlog hit 1.5σ. 2. **The Trial Expiry Nurture workflow alone accounts for 142 of 222 errors (64%).** Single-cause spikes are easy to fix; multi-cause spikes (errors evenly distributed across workflows) usually mean a portal-wide change (auth refresh, integration timeout, bulk-import overrun). 3. **The "Welcome step 2" unpublish (58 errors) is a different cause but coincident.** A marketer was redesigning the welcome series and accidentally unpublished a step. Vortex IQ shows both as contributors. The workflow does not pause; it errors on each enrolment until the step is republished. 4. **Steady-state errors (22) are normal noise.** Each portal has a baseline of INVALID\_EMAIL\_ADDRESS errors from typo'd form submissions; these are expected and contribute to the 7-day mean. They do not cause spikes; only deviations above baseline do. 5. **Recovery is automatic once the cause is fixed.** Re-creating `trial_end_date` (or pointing the workflow at a renamed equivalent) and re-publishing the welcome step both stop new errors. The 24h trailing count decays over the next 24 hours and falls below the threshold; the alert clears on the next evaluation cycle. ## Sibling cards merchants should reference together Workflow Error Spike is the leading indicator. Pair with downstream cards to confirm impact: | Card | Why pair it with Workflow Error Spike | | ---------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------ | | [Workflows With Errors](/nerve-centre/kpi-cards/hubspot/workflows-with-errors) | The 7-day rolling list view. This alert fires on change; that card lists ongoing problems. | | [Marketing Emails Sent](/nerve-centre/kpi-cards/hubspot/marketing-emails-sent) | Send-impacting errors (UNPUBLISHED\_EMAIL, SEND\_FROM\_NOT\_VERIFIED) collapse send volume; pair to confirm. | | [Active Workflows](/nerve-centre/kpi-cards/hubspot/active-workflows) | The denominator. 222 errors across 38 workflows is 5.8 per workflow per day; healthy is below 2. | | [Workflow Enrolment Volume](/nerve-centre/kpi-cards/hubspot/workflow-enrolments) | Enrolment-step errors block downstream workflow steps; an error spike often coincides with an enrolment plateau or drop. | | [Email Deliverability Drop](/nerve-centre/kpi-cards/hubspot/email-deliverability-drop) | Adjacent operational alert. Send-from-address errors here often coincide with deliverability drops next-day. | | [Pipeline Drop Alert](/nerve-centre/kpi-cards/hubspot/pipeline-drop-alert) | Sometimes related: a workflow that auto-creates deals breaks, downstream pipeline contraction follows. | | [Form Submission Volume](/nerve-centre/kpi-cards/hubspot/form-submissions) | Form-trigger workflows fail when forms are renamed; pair to confirm whether the upstream chain broke. | | [Workflow-Triggered Commerce Revenue](/nerve-centre/kpi-cards/hubspot/workflow-triggered-commerce-revenue) | The cross-platform revenue impact, do the broken workflows drive commerce \$. | ## Reconciling against the vendor's own dashboard **Where to look in HubSpot:** Two native views cover the spike's components: > [HubSpot → Automation → Workflows](https://app.hubspot.com/workflows) (filter to "Errors" facet, sorted by recency) > [HubSpot → Automation → Workflows → individual workflow → History tab](https://app.hubspot.com/workflows) (per-workflow event log) The error-recency facet should show the same workflows in this card. **Why our number may legitimately differ from HubSpot's:** | Reason | Direction | Why | | ------------------------- | ------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------- | | **Sync lag** | Ours lower for the most recent 5-15 minutes | Errors fire instantly in HubSpot; we trail by one sync cycle. | | **Window definition** | Either | This card uses a 24h rolling window; HubSpot's UI typically shows "last 7 days" or "all time" per workflow. The aggregate counts differ. | | **Stale-error retention** | Either | HubSpot retains workflow error history for 30+ days; we keep only what is needed for the 24h alert plus the 7-day list. | | **Workflow scope** | Ours wider | HubSpot's UI shows workflows the user has access to. Vortex IQ reads through portal-wide token. | | **Threshold tuning** | n/a | HubSpot does not have an equivalent spike-detection alert; the comparison is against the raw error log, not against a baseline. | **Cross-connector reconciliation:** | Card | Expected relationship | What causes legitimate divergence | | ------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------- | | [Klaviyo Flow Errors (when added)](/nerve-centre/klaviyo/klv_flow_errors) | If a HubSpot workflow webhook drives a Klaviyo flow, errors here often correlate with Klaviyo flow drop-offs | Direct API calls bypass workflows entirely; webhook retries may mask the error. | | [GA4 Email channel sessions](/nerve-centre/google_analytics/ga_email_sessions) | Send-impacting errors (UNPUBLISHED\_EMAIL) reduce email-channel sessions within 24-48 hours | UTM-tagging gaps may understate the channel impact. | ***
Documentation cross-reference (for agencies running multiple automation platforms) Workflow error spike detection is universal in marketing-automation. Same shape on: * [`klaviyo.klv_flow_error_spike`](/nerve-centre/klaviyo/klv_flow_error_spike) (when added) * [`mailchimp.mc_automation_error_spike`](/nerve-centre/mailchimp/mc_automation_error_spike) (when added) The 2σ-vs-baseline approach is the cross-platform standard.
## Known limitations / merchant FAQs **The alert fired but the dashboard says zero workflows in error, why?** The alert fires on a 24h spike; by the time the merchant looks, errors may have been fixed. Check the audit log for the original event time. The Workflows With Errors card retains the rolling 7-day list for follow-up. **What is the difference between this card and Workflows With Errors?** This card detects sudden change (a spike vs baseline). Workflows With Errors is the persistent list (any errored workflow in the last 7 days). Use both: the alert tells you something just broke; the list tells you what is still broken. **Can I tune the 2σ threshold?** Yes, configurable per portal. Tighter thresholds (1.5σ) catch more anomalies but increase false positive rate. Looser (2.5σ) only catches major events. 2σ is the default for most B2B portals. **Why does the alert clear before I have time to act?** If the underlying error is fixed quickly, the 24h trailing count decays below the threshold. We do not auto-suppress; the audit log retains the alert fire-and-clear sequence. For long-running incidents, the alert stays active until the cause is fully resolved. **INVALID\_EMAIL\_ADDRESS spikes, what do they mean?** A campaign or import dumped many bad emails into a workflow. The workflow itself is fine; the data is not. Fix is upstream form validation or import-deduplication. **RATE\_LIMIT\_EXCEEDED, what causes that?** HubSpot's automation API has per-portal rate limits. Bulk operations (mass imports, mass workflow re-enrolment) can spike errors briefly. Usually self-resolves within an hour. **Multi-portal aggregation, can I see spikes across portals?** No. One card per portal because the 7-day baseline is portal-specific. Aggregating across portals would dilute high-signal alerts on small portals. **List-segment lag, can it cause a spike?** Indirectly. A list-trigger workflow that mass-enrols based on list membership can cause a brief RATE\_LIMIT spike during the enrolment-batch window. Usually self-resolves. **Lifecycle-stage backfill, does it touch this?** Lifecycle backfills that trigger downstream workflows (e.g. MQL handoff) can cause transient spikes. The alert may fire during the backfill window and clear once it completes. Audit HS01 surfaces backfill-triggered spikes separately. **Today-volatility, why is the count higher in business hours?** Workflows enrol contacts most actively during business hours (forms submitted, sales activities logged). Errors scale with enrolment volume. The 7-day baseline captures the bimodal hourly pattern, so business-hour error counts are expected. *** ### Tracked live in Vortex IQ Nerve Centre *Workflow Error Spike* is one of hundreds of KPI pulses Vortex IQ tracks across HubSpot and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English. [Start for free](https://app.vortexiq.ai/login) or [book a demo](https://www.vortexiq.ai/contact-us) to see this metric running on your own data.