Dashboards - KaireonAI

KaireonAI ships with a suite of monitoring dashboards, each designed for a different audience and set of questions. Together they give you real-time visibility into every layer of the platform — from infrastructure health to business outcomes to model drift. All dashboards auto-refresh every 30 seconds. This page covers the five operational dashboards below; the Executive dashboard is documented on its own page. Navigate to Dashboards in the sidebar to access them.

Dashboard	Audience	Questions It Answers
Operations	Platform engineers	Is the system healthy? How fast are decisions? Is the DLQ growing?
Business	Analysts, marketers	Which offers convert? Which channels perform? What is the funnel?
Data Health	Data engineers	Are connectors up? Are pipelines running? How many schemas exist?
Model Health	Data scientists	Is model accuracy stable? Which features matter? Are experiments running?
Attribution	Growth, analytics	Which channels drive conversions? How should credit be split?

Operations Dashboard

Path: /dashboards/operations The Operations Dashboard is the primary system-health view for platform engineers. It answers: Is the decisioning engine running fast and correctly?

Performance KPI Cards

Four cards across the top row:

Card	What It Shows	Source
Total Runs	Count of all Decision Flow executions	Runs API
Avg Latency	Mean latency in milliseconds	Runs API
P95 Latency	95th-percentile latency	Runs API
Success Rate	`completed / (completed + failed)` as percentage	Runs API

Latency Visualizations

Chart	Description
Latency Distribution	Bar chart where each bar = one run, height proportional to latency vs. P95. Hover for run ID and exact latency.
Latency Trend (7-day)	Line chart with P50 and P99 series over the past week. Shows sample data badge when no real time-series data exists.

Acceptance Rate by Offer

Bar chart plus detail table showing per-offer acceptance rates from the interaction_summaries table:

Column	Description
Offer ID	Unique identifier
Impressions	Total presentations
Positive	Accepted/converted interactions
Negative	Rejected/dismissed interactions
Accept Rate	`positive / impressions` as percentage

Budget Utilization

Progress bars for offers with budget allocations. Color-coded:

Color	Condition	Meaning
Green	Below 70%	Healthy spend rate
Amber	70—90%	Approaching limit
Red	Above 90%	Near or at budget exhaustion

Decision Pipeline Panel

Reads persisted DecisionTrace aggregates via GET /api/v1/dashboards/decision-pipeline. The window is driven by the page’s period selector and capped at 168 hours (7 days); the panel header shows the effective window (“last 7d”, “last 24h”, etc.) derived from the API response’s windowHours field — so it is always truthful. Query param: windowHours (1–168, default 24). The period selector now correctly refetches the panel on change.

Field	Description
Duration P50 / P95 / P99	Percentile aggregates of `totalLatencyMs` across all traces in the window
Funnel	Average candidate counts at 5 stages: Initial Candidates, After Qualification, After Suppression, After Contact Policy, After Scoring
Filter rates	Cumulative fraction of initial candidates dropped — `qualification` = (initial − after_qual) / initial; `contactPolicy` = (initial − after_contact) / initial. Both are cumulative (not stage-isolated), shown as progress bars (red > 80%, amber > 50%)

When traceCount === 0 the panel shows an empty state. The empty state note mentions decisionTraceSampleRate (configurable under Settings → Tenant) in case tracing is turned down or off. The panel loads via a direct fetch that respects the page’s period selector (periodDays → windowHours = min(periodDays × 24, 168)); auto-refreshes every 30 seconds alongside the rest of the dashboard.

Dead Letter Queue (DLQ) Panel

Status	Condition	Color
Healthy	0 events	Green
Warning	1—10 events	Amber
Critical	10+ events	Red

Shows topic breakdown (e.g., decision.outcome, pipeline.error) with Retry All and Purge action buttons.

Circuit Breaker Panel

Displays state change history for each breaker (e.g., enrichment-redis, connector-snowflake). Color-coded badges: red for open, amber for half_open, green for closed.

Investigating a Drop in Acceptance Rate

Suppose you notice offer_summer_promo dropped from 12% to 3% acceptance overnight. Here is how to investigate:

Check the Decision Pipeline panel. If “After Qualification” is normal but “After Contact Policy” drops sharply, a new frequency cap is filtering aggressively.
Open a Decision Trace. The trace shows candidates at each stage. If the offer is present at Qualification but absent after Contact Policy, expand the contact policy section to see which rule suppressed it.
Check the Filter Rate bars. Contact Policy filter rate above 80% (red) means most candidates are being suppressed — likely a misconfigured policy.

If DLQ depth rises while acceptance rates drop, check whether the outcome recording pipeline is failing. Missing outcomes make acceptance rate calculations unreliable.

Business Dashboard

Path: /dashboards/business The Business Dashboard answers: How are my offers performing? Designed for analysts and marketers who need to track conversion rates, channel effectiveness, and revenue.

Summary Cards

Card	Source
Active Offers	`Offer` records with `status = "active"`
Active Channels	`Channel` records with `status = "active"`
Experiments	Active `Experiment` records
Journeys	Total `Journey` records
Active Triggers	Active trigger rule records
Pending Approvals	Pending approval request records
Acceptance Rate	Aggregate positive / total interactions across all offers (last 30 days, from the `interaction_summaries` table)

Offer Funnel

Four-stage funnel showing conversion from configuration to delivery readiness:

Total Offers — all offers in the tenant
Active — offers with status = "active"
With Creatives — active offers that have at least one active creative
Active Creatives — total count of active creatives

If you have 20 active offers but only 5 have creatives, 15 offers cannot be delivered. The funnel makes this gap immediately visible.

Charts

Chart	What It Shows
Channel Effectiveness (pie)	Each channel’s share of conversions (`positive / impressions * 100`)
Offer Performance (bar)	Horizontal bars ranking offers by conversion count
Daily Trend (line, 7-day)	Impressions and conversions over time from interaction history

Offer Performance Detail Table

Column	Description
Offer	Offer name
Priority	Configured priority value
Creatives	Count of active creatives
Impressions	Total presentations
Conversions	Positive interactions
Conv. Rate	`conversions / impressions * 100`
Revenue	Sum of `totalValue` from interaction summaries

High impressions but zero conversions? Check if the creative is compelling and the channel matches the audience. Conversions but zero revenue? Verify that Respond API calls include a value field.

Data Health Dashboard

Path: /dashboards/data-health Monitors the ingestion layer — connectors, pipelines, and schemas.

Summary Cards

Card	Description
Total Connectors	All configured connectors (S3, Kafka, Snowflake, etc.)
Active Connectors	Connectors with `status = "active"`
Pipelines	Total pipeline definitions
Schemas	Total data schema definitions

Connectors Table

Column	Description
Name	Connector name
Type	e.g., AWS S3, Snowflake, Kafka
Status	`active` (green), `error` (red), `inactive` (gray)
Auth	Authentication method (e.g., `iam_role`, `api_key`, `oauth2`)

A connector showing error usually means the last connection test failed. Navigate to Data > Connectors, select it, and run Test Connection for the specific error. Common causes: expired credentials, changed IP allowlists, network policy changes.

Model Health Dashboard

Path: /dashboards/model-health Tracks scoring model status, accuracy trends, and experiment activity. Designed for data scientists monitoring model performance.

Summary Cards

Card	Description
Total Models	All registered algorithm models
Active Models	Models with `status = "active"`
Experiments	Total experiment definitions

Charts

Chart	What It Shows
AUC Trend (line, 7-day)	Area Under Curve over time for the primary active model
Feature Importance (bar)	Top predictors ranked by importance weight

Models Table

Column	Description
Name	Model name
Type	Scorecard, Bayesian, or Gradient Boosted
Status	`active`, `error`, or `draft`
AUC	Latest AUC metric as percentage

A sudden AUC drop (more than 5 points in a week) suggests feature drift. Check the Feature Importance chart for ranking changes, then consider retraining with recent data.

Analysis panels

Four additional analysis panels sit below the main cards:

Uplift

Source: GET /api/v1/treatments/uplift?windowDays=N Per-creative observational uplift cards. Window selector: 7 / 30 / 90 days; auto-refetches on change. Renders a bar chart of uplift value per creative and a table with sample counts, p-value, and 95% confidence interval (unpooled Wald standard error). A samplesBelowMinimum badge flags creatives without enough data for reliable inference.

CATE Explorer

Source: GET /api/v1/algorithm-models/[id]/uplift?customerId=...&method=... Submit-driven panel for per-customer Conditional Average Treatment Effect (τ). Pick a model, enter a customerId, select a method (t_learner or x_learner), and submit. Renders an offers table with τ values, color-coded segment badges (persuadable / sure_thing / lost_cause / sleeping_dog / uncertain), and an overall ATE summary line.

Fairness

Source: POST /api/v1/fairness/evaluate with { mode: "inline", samples: [...] } Inline-mode fairness evaluation. Paste a JSON array of { group: string, prediction: boolean, label?: boolean } decision samples and submit. Renders per-group stats (count, positive rate, TPR/FPR when labels are present), key metric cards (demographic parity gap, disparate impact ratio, four-fifths badge), and a warnings list.

There is no persisted fairness history table. The fairness-recheck cron writes outcomes to AuditLog entries but does not store structured FairnessEvaluation rows. The panel is therefore form-driven and stateless — it evaluates samples you provide inline. To see historical gate outcomes, query GET /api/v1/audit-logs?action=fairness_evaluate.

Drift Check

Source: POST /api/v1/models/[id]/drift Interactive drift analysis. Select a model, paste JSON for reference and current feature distributions ({ featureName: number[] }), and submit. Renders a per-feature PSI and KS table with severity badges (none / monitor / alert) and an overall verdict.

Attribution Dashboard

Path: /dashboards/attribution Multi-touch attribution analysis comparing how different models distribute credit across channels. Helps answer: Which channels actually drive conversions, and how should I allocate budget?

Attribution Models

Five models are selectable. Two additional models appear as disabled Coming Soon entries in the picker — they are rendered but not selectable and the API enum is unchanged.

Model	How Credit Is Assigned	Status
Last Touch	100% to the final touchpoint before conversion	Live
First Touch	100% to the first touchpoint	Live
Linear	Equal split across all touchpoints	Live
Time Decay	More credit to touchpoints closer to conversion	Live
Position Based	40% first, 40% last, 20% split across middle	Live
Shapley	Game-theoretic credit assignment across touchpoints	Coming soon — not selectable
Cross-Device	Identity-stitched attribution across devices	Coming soon — not selectable

Summary KPIs

Card	Description
Total Conversions	Number of attribution results
Channels	Distinct channels across all touchpoints
Total Credit	Sum of all credit values
Avg Touchpoints / Conversion	Mean touchpoints per conversion journey

Channel Contribution Chart

Horizontal bar chart showing each channel’s percentage contribution to total credit, with raw credit values.

Conversions Per Channel Table

Column	Description
Channel	Channel identifier
Touchpoints	Interaction count for this channel
Total Credit	Sum of credit assigned
Contribution %	`channel credit / total credit * 100`

Compare the same data across attribution models. If a channel ranks number 1 under Last Touch but number 4 under First Touch, it is strong at closing but weak at introducing — useful for budget allocation decisions.

Prometheus Metrics

KaireonAI exposes a Prometheus-compatible scrape endpoint at /api/metrics (requires admin role).

Scrape Configuration

scrape_configs:
  - job_name: "kaireon"
    scheme: https
    metrics_path: /api/metrics
    authorization:
      type: Bearer
      credentials: "<your-admin-api-key>"
    static_configs:
      - targets: ["your-kaireon-instance.com"]
    scrape_interval: 15s

Key Metrics

Decision Engine

Metric	Type	Description
`kaireon_decision_latency_ms`	Histogram	Decision engine latency (by channel)
`kaireon_decision_pipeline_duration_ms`	Histogram	Full pipeline execution time
`kaireon_decision_candidates`	Gauge	Candidate count at each stage
`kaireon_decision_qualification_filter_rate`	Gauge	Qualification filter ratio
`kaireon_decision_contact_policy_filter_rate`	Gauge	Contact policy filter ratio
`kaireon_decision_delivery_total`	Counter	Total responses delivered
`kaireon_qualification_unknown_rule_total`	Counter	Unknown decisioning gate types — drift signal between code and tenant config (by `ruleType`)

HTTP & API

Metric	Type	Description
`kaireon_http_request_duration_seconds`	Histogram	HTTP request duration (by method, route, status)
`kaireon_http_error_total`	Counter	HTTP 4xx/5xx errors

Pipelines & Connectors

Metric	Type	Description
`kaireon_pipeline_execution_latency_ms`	Histogram	Pipeline execution time
`kaireon_pipeline_rows_processed_total`	Counter	Total rows processed
`kaireon_connector_test_latency_ms`	Histogram	Connector test duration
`kaireon_connector_test_total`	Counter	Connector test attempts (by `connectorType`, `status`) — pair with the latency histogram for failure-ratio alerting

Infrastructure

Metric	Type	Description
`kaireon_cache_hits_total` / `kaireon_cache_misses_total`	Counter	Cache performance
`kaireon_dlq_depth`	Gauge	Dead letter queue depth (by tenant)
`kaireon_circuit_breaker_state_change_total`	Counter	Circuit breaker transitions
`kaireon_active_worker_jobs`	Gauge	Active background jobs
`kaireon_outbox_pending_count`	Gauge	Per-tenant outbox backlog. Refreshed by the outbox processor on every poll tick. See Outbox publisher
`kaireon_identity_resolution_total`	Counter	Identity-graph resolution attempts (by `result`: `matched` / `created` / `ambiguous` / `error`)

Experiments & Models

Metric	Type	Description
`kaireon_experiment_assignment_total`	Counter	Experiment variant assignments
`kaireon_scoring_model_failure_total`	Counter	Scoring model failures (fallback triggered)
`kaireon_mandatory_cap_hit_total`	Counter	Mandatory offer daily cap hits

AI Intelligence

Metric	Type	Description
`kaireon_ai_intelligence_calls_total`	Counter	AI tool invocations (by `tool`, `status`) — covers MCP intelligence tools, in-app AI assistant, and intelligence dashboard
`kaireon_ai_intelligence_duration_seconds`	Histogram	AI tool execution duration (by `tool`); buckets 0.1, 0.5, 1, 2, 5, 10 seconds

Compliance

Metric	Type	Description
`kaireon_gdpr_erasure_total`	Counter	GDPR erasure requests
`kaireon_dsar_request_total`	Counter	DSAR requests (by type, status)
`kaireon_sso_auth_total`	Counter	SSO auth attempts (by provider, result)

The /api/v1/metrics/summary endpoint returns a curated subset of these metrics as JSON. This is what the Operations Dashboard uses internally — it does NOT include every metric on this page; the full scrape feed lives at /api/metrics. For the full operator-grade reference of the seven post-W10 metrics — including registered file-line, alert guidance, and PromQL expressions — see Metrics Reference.

Segment Dimension

Six of the Business and Operations dashboard queries accept an optional segmentId query parameter to scope their aggregation to customers in a specific segment. When provided, the underlying SQL joins the segment’s materialized view (seg_<id_prefix>) onto the interaction tables by customerId.

Dashboard data `type=`	Segment filter behavior
`acceptance_rate`	Per-offer acceptance scoped to the segment.
`offer_performance`	Top-20 offer impressions/conversions/revenue within the segment.
`offer_performance_grouped`	`groupBy=segment` fans the response out by every active materialized segment; each offer row is decorated with `segmentId` + `segmentName`. Alternatively, pass `segmentId=<id>` with `groupBy=channel` or `groupBy=category` to filter within one segment.
`channel_effectiveness`	Channel-level impressions/conversions scoped to the segment.
`daily_trend`	7-day impression/conversion line scoped to the segment.
`revenue_trend`	N-day revenue line from interaction summaries scoped to the segment.

Behavior when a segment is not ready

Segments are materialized to PostgreSQL views asynchronously. If a segment exists but its view has not been built (for example a draft segment), the endpoint returns {"data": [], "warning": "segment view not materialized"} — dashboards can render an informational banner instead of an error. An unknown segmentId returns {"data": [], "warning": "unknown segmentId"} under the same pattern.

Example

curl -H "X-API-Key: $API_KEY" -H "X-Tenant-Id: $TENANT_ID" \
  "https://playground.kaireonai.com/api/v1/dashboard-data?type=offer_performance&segmentId=seg_vip&days=7"

See the Dashboard Data API reference for complete parameter and response definitions.

Export and Save as Report

Every dashboard in the platform ships with two header-bar buttons that turn the current view into a report artefact:

Export works unconditionally — no cron wiring required. Clicking Export renders the artifact server-side and streams the file back to the browser immediately.Save as Report also creates the template + schedule immediately, but the resulting saved schedule only fires on cadence once /api/cron/tick is being invoked by an external scheduler. During pilot / initial deployment this is usually not wired. Use the Run Now button on the saved template in /settings/reports for on-demand delivery until you follow EventBridge Setup (optional).

Export

Dropdown with PDF / CSV / Markdown / HTML options. Clicking a format:

Builds a transient report template from the dashboard’s current filters (date range, segment, etc.) via the view-to-template.ts bridge.
POSTs the transient template to /api/v1/reports/preview — no database row is created.
Receives a base64-encoded artifact and triggers a browser download via Blob + URL.createObjectURL.

CSV exports skip LLM narration to keep latency low; PDF / Markdown / HTML include a narrated executive summary when an AI provider is configured.

Save as Scheduled Report

Opens a modal pre-filled from the current view. Pick:

Name (required; defaults to {Dashboard label} · {days}d).
Formats (multi-select; default PDF).
Narrative toggle.
Schedule (preset: Daily 8am / Weekly Mon 8am / Monthly 1st 8am / Custom cron).
Destinations (multi-select of configured notification providers — Slack, Teams, webhook, Ops-email).

Submitting creates a report template and a report schedule. When the cron is wired (AWS EventBridge → /api/cron/tick), the schedule runs on cadence and delivers artefacts to every destination. Until then, trigger delivery from /settings/reports using Run Now, or invoke the run-now endpoint at /api/v1/reports/templates/[id]/run-now directly. View the persisted template at /settings/reports.

Which data sources are sent?

Each dashboard declares its source list in the dashboards-to-report-template bridge:

Dashboard	Data sources
Executive	`offer_performance`, `channel_effectiveness`, `revenue_trend`, `anomaly_candidates`, `selection_frequency`
Business	`offer_performance`, `channel_effectiveness`, `revenue_trend`, `funnel`
Operations	`daily_trend`, `decision_traces_summary`
Model Health	`selection_frequency`, `anomaly_candidates`
Data Health	`daily_trend`
Attribution	`offer_performance`, `channel_effectiveness`

Every source is tenant-scoped — the API enforces tenant context and filters every database query by tenant.

Executive Dashboard

C-suite summary with narrated weekly highlights and KPI deltas.

Reports

Templates, schedules, formats, and delivery — the engine behind Save-as-Report.

Decision Traces

Forensic tracing for debugging qualification and ranking.

Metrics Reference

Operator-grade reference for the post-W10 metrics with PromQL alert expressions.

Algorithms & Models

Understand the scoring models tracked by Model Health.

​Operations Dashboard

​Performance KPI Cards

​Latency Visualizations

​Acceptance Rate by Offer

​Budget Utilization

​Decision Pipeline Panel

​Dead Letter Queue (DLQ) Panel

​Circuit Breaker Panel

​Investigating a Drop in Acceptance Rate

​Business Dashboard

​Summary Cards

​Offer Funnel

​Charts

​Offer Performance Detail Table

​Data Health Dashboard

​Summary Cards

​Connectors Table

​Model Health Dashboard

​Summary Cards

​Charts

​Models Table

​Analysis panels

​Uplift

​CATE Explorer

​Fairness

​Drift Check

​Attribution Dashboard

​Attribution Models

​Summary KPIs

​Channel Contribution Chart

​Conversions Per Channel Table

​Prometheus Metrics

​Scrape Configuration

​Key Metrics

​Segment Dimension

​Behavior when a segment is not ready

​Example

​Export and Save as Report

​Export

​Save as Scheduled Report

​Which data sources are sent?

​Related

Executive Dashboard

Reports

Decision Traces

Metrics Reference

Algorithms & Models

Operations Dashboard

Performance KPI Cards

Latency Visualizations

Acceptance Rate by Offer

Budget Utilization

Decision Pipeline Panel

Dead Letter Queue (DLQ) Panel

Circuit Breaker Panel

Investigating a Drop in Acceptance Rate

Business Dashboard

Summary Cards

Offer Funnel

Charts

Offer Performance Detail Table

Data Health Dashboard

Summary Cards

Connectors Table

Model Health Dashboard

Summary Cards

Charts

Models Table

Analysis panels

Uplift

CATE Explorer

Fairness

Drift Check

Attribution Dashboard

Attribution Models

Summary KPIs

Channel Contribution Chart

Conversions Per Channel Table

Prometheus Metrics

Scrape Configuration

Key Metrics

Segment Dimension

Behavior when a segment is not ready

Example

Export and Save as Report

Export

Save as Scheduled Report

Which data sources are sent?

Related