Analyzing operational data...

Scanning traces for automation opportunities

Operational Intelligence

Inbound → operator actions → system events → resolution

orders · operators · last 30 days

Hover any metric for the ClickHouse query powering it

Orders Tracked
Total unique order workflows observed in the trace data.
Count of unique trace IDs where an order was created in the last 30 days.

Powered by: uniqExact(TraceId)

SELECT uniqExact(TraceId) FROM otel_traces WHERE ServiceName='clickstack-portal' AND SpanName='order.created' AND Timestamp >= now()-INTERVAL 30 DAY

Celonis equivalent: Process event count

Orders Tracked

order lifecycles

Operators Observed
Distinct operators with at least one automation-candidate task.
Sourced from the /api/opportunities scan over the last 7 days.

Powered by: uniqExact()

SELECT uniqExact(SpanAttributes['operator.id']) AS operators_affected FROM otel_traces WHERE SpanAttributes['automation.candidate'] = 'true' AND Timestamp >= now() - INTERVAL 7 DAY

Celonis equivalent: Resource analysis

Operators Observed

unique operators

Avg Resolution Time
Mean time from first event to last event per order, in days.
Average elapsed days from first to last event per order trace.

Powered by: dateDiff() + avg()

SELECT avg(dateDiff('hour', min_ts, max_ts)) / 24 FROM ( SELECT TraceId, min(Timestamp) as min_ts, max(Timestamp) as max_ts FROM otel_traces WHERE ServiceName='clickstack-portal' GROUP BY TraceId HAVING max_ts > min_ts )

Celonis equivalent: Throughput time

Avg Resolution Time

inbound to close

Portal Pages Captured
Total page navigation events recorded by HyperDX.
Browser page visits with DOM metadata (element count, scroll depth, content).

Powered by: count()

SELECT count() FROM otel_traces WHERE ServiceName='clickstack-portal' AND SpanName LIKE 'page.%' AND Timestamp >= now()-INTERVAL 30 DAY

No Celonis equivalent — this is unique to ClickStack

Portal Pages Captured

portal layer

Conformance
Percentage of orders that follow the designed "happy path" — the intended step sequence with no deviations, rework, or escalations.
% of orders completing the full 5-step happy path sequence.

Powered by: windowFunnel()

SELECT round(countIf(level >= 5) / count() * 100, 1) FROM ( SELECT TraceId, windowFunnel(2592000)(Timestamp, SpanName='order.created', SpanName='inspection.scheduled', SpanName='report.submitted', SpanName='review.completed', SpanName='order.closed' ) as level FROM otel_traces WHERE ServiceName='clickstack-portal' GROUP BY TraceId )

Celonis equivalent: Conformance Checker

Conformance

happy path %

Rework Rate
Percentage of orders that contain at least one revision loop — where a completed step was sent back for correction.
% of orders with revision bounce loops (review → revision needed).

Powered by: countIf()

SELECT round(countIf(rework > 0) / count() * 100, 1) FROM ( SELECT TraceId, countIf(SpanName IN ( 'review.revision_needed', 'report.revised', 'status.revision_needed' )) as rework FROM otel_traces WHERE ServiceName='clickstack-portal' GROUP BY TraceId )

Celonis equivalent: Rework analysis

Rework Rate

revision loops

Activity · 7 days

Orders Today
Distinct order traces created since midnight

Powered by: count(DISTINCT TraceId)

SELECT count(DISTINCT TraceId) FROM otel_traces WHERE ServiceName='clickstack-portal' AND SpanName='order.created' AND Timestamp >= today()

Celonis equivalent: included in base Process Mining license

Orders Today --

Avg Duration
Average elapsed time from first to last event per order

Powered by: formatReadableTimeDelta()

SELECT formatReadableTimeDelta(avg(dur)) FROM ( SELECT TraceId, dateDiff('second', min(Timestamp), max(Timestamp)) as dur FROM otel_traces GROUP BY TraceId HAVING dur > 0 )

Celonis equivalent: Throughput Time

Avg Duration --

Events/min
Event throughput rate over the last hour

Powered by: count() / dateDiff()

SELECT round(count() / dateDiff('minute', min(Timestamp), max(Timestamp)), 1) FROM otel_traces WHERE Timestamp >= now()-INTERVAL 1 HOUR

Celonis equivalent: Process Throughput

Events/min --

Error Rate
Percentage of spans with ERROR status in the last 24 hours

Powered by: countIf() / count()

SELECT round( countIf(StatusCode='ERROR') / count() * 100, 2) FROM otel_traces WHERE ServiceName='clickstack-portal' AND Timestamp >= now()-INTERVAL 24 HOUR

Celonis equivalent: Process Health Monitoring

Error Rate --

Identified Automation Targets
Tasks ranked by weekly hours recoverable through automation

Powered by: count() + avg(Duration)

SELECT SpanName as task, count() / 7 as daily_freq, avg(Duration/1e9) / 60 as avg_minutes, round(count() * avg(Duration/1e9) / 3600, 1) as weekly_hours FROM otel_traces WHERE ServiceName='clickstack-portal' AND Timestamp >= now()-INTERVAL 7 DAY GROUP BY task ORDER BY weekly_hours DESC

Celonis equivalent: Task Mining

Identified Automation Targets

Tasks ranked by weekly hours spent. High-frequency, low-complexity tasks are the best automation candidates.

CONNECTING…

UPDATED: --:--:--

Object-Centric Process Mining
Multi-object interaction analysis: tracks how different object types (emails, portal pages, API calls, inspections) interact within each order

Powered by: ClickHouse-native OCPM

-- Full OCPM pipeline via /api/ocpm -- Interaction matrix, lifecycle durations, -- bottleneck detection, and operator affinity -- all computed in standard ClickHouse SQL

Celonis equivalent: OCPM Module

Object-Centric Process Mining ? OCPM analyzes how different object types (orders, vendors, inspectors, documents) interact within business processes. Unlike traditional process mining which follows one case ID, OCPM tracks multiple objects simultaneously — revealing cross-functional bottlenecks that single-case analysis misses.

Learn more →

Multi-object interaction analysis across 8 span categories. First ClickHouse-native OCPM implementation.

OCPM LIVE

Hover any metric for the ClickHouse query powering it

Object Categories
Distinct span prefix categories with activity in this period

Powered by: multiIf() + uniqExact()

SELECT uniqExact(multiIf( SpanName LIKE 'email.%', 'Communication', SpanName LIKE 'page.%', 'Portal', SpanName LIKE 'api.%', 'System', SpanName LIKE 'inspection.%', 'Field Work', 'Other')) FROM otel_traces WHERE ServiceName='clickstack-portal'

Celonis equivalent: Object Type Registry — OCPM add-on

Object Categories

active types

Cross-Object Interactions
Unique category pairs that co-occur within the same order

Powered by: arrayJoin() + CROSS JOIN

SELECT count(DISTINCT (cat_a, cat_b)) FROM ( SELECT arrayJoin(categories) AS cat_a, arrayJoin(categories) AS cat_b FROM order_categories WHERE cat_a < cat_b ) -- Derived from /api/ocpm logic

Celonis equivalent: Object Interaction Graph — OCPM add-on

Interactions

category pairs

Slowest Cross-Object Transition
The category transition with the highest average duration

Powered by: lagInFrame() + avg()

SELECT cat_from, cat_to, avg(time_between_ms) as avg_handoff_ms FROM object_transitions GROUP BY cat_from, cat_to ORDER BY avg_handoff_ms DESC LIMIT 1 -- Derived from /api/ocpm logic

Celonis equivalent: Cross-Object Bottleneck — requires OCPM add-on

Slowest Handoff

cross-category

Path Variants
Total unique path patterns across all object categories

Powered by: cityHash64() + uniqExact()

SELECT uniqExact(path_hash) FROM ( SELECT TraceId, groupArray(SpanName) as path, cityHash64(toString( groupArray(SpanName))) as path_hash FROM otel_traces WHERE ServiceName='clickstack-portal' GROUP BY TraceId )

Celonis equivalent: Process Variants by Object Type — PQL required

Total Variants

unique patterns

Object Interaction Matrix
Category co-occurrence counts across order traces

Powered by: arrayJoin() + count()

SELECT cat_from, cat_to, count() as transitions, avg(time_between_ms) as avg_handoff_ms FROM object_transitions GROUP BY cat_from, cat_to ORDER BY transitions DESC -- Derived from /api/ocpm

Celonis equivalent: OCPM Module

Object Interaction Matrix

Object Lifecycle Duration
How long each object category stays active per order

Powered by: dateDiff() + quantile()

SELECT category, count() as order_count, avg(duration_hours) as avg_hours, median(duration_hours) as median_hours, quantile(0.95)(duration_hours) as p95_hours FROM category_lifecycles GROUP BY category ORDER BY avg_hours DESC

Celonis equivalent: Object Lifecycle Analysis — OCPM add-on

Object Lifecycle Duration

Cross-Object Bottlenecks
Slowest transitions between object categories by P95 latency

Powered by: quantile(0.95)()

SELECT cat_from, cat_to, quantile(0.95)(time_between_ms) / 3600000 as p95_hours FROM object_transitions GROUP BY cat_from, cat_to ORDER BY p95_hours DESC LIMIT 5

Celonis equivalent: Cross-Object Bottleneck Analysis — OCPM add-on

Cross-Object Bottlenecks

Operator-Object Affinity
Which operators interact most with each object category

Powered by: multiIf() + count()

SELECT SpanAttributes['operator.id'] as operator, multiIf( SpanName LIKE 'email.%', 'Communication', SpanName LIKE 'page.%', 'Portal', SpanName LIKE 'api.%', 'System', SpanName LIKE 'inspection.%', 'Field Work', 'Other') as category, count() as interactions FROM otel_traces WHERE ServiceName='clickstack-portal' GROUP BY operator, category ORDER BY operator, interactions DESC

Celonis equivalent: Resource Analysis

Operator-Object Affinity

A ClickHouse-native OCPM implementation built on standard SQL — complements existing process-mining tooling without a separate platform.

Dashboard powered by synthetic demonstration data. Production metrics will differ.

Response-Determinism Analysis

score = templateConc x auto-resolve x confidence x completeness

ClickHouse Execution Trace

-- Loading...

Operational Intelligence

Identified Automation Targets

Top Operational Flows

Object Interaction Matrix

Object Lifecycle Duration

Cross-Object Bottlenecks

Operator-Object Affinity

Keyboard Shortcuts

Node Details

Operational Intelligence

Identified Automation Targets

Top Operational Flows

Object Interaction Matrix

Object Lifecycle Duration

Cross-Object Bottlenecks

Operator-Object Affinity

PATH_ANALYSIS

ClickHouse Execution Trace

Node Details