Tealfabric

Operations triage

Dashboard → Monitor → queue → trace → resolve

01 Dashboard02 Monitor03 Queue04 Trace05 Resolve

Home dashboard · tenant acme-ops · 06:42

Failed process runs (2)

nightly_erp_sync · VALIDATION_ERROR · 02:14

vendor_intake_submit · TIMEOUT · 04:55

Unread notifications (3)

Integration rate limit warning · hubspot-marketing

agent todos: 1 open · quick link → Monitor

Simulated flow — nightly sync failure triaged from dashboard through Monitor to fix and re-run

Built to be operated

“Build in Tealfabric. Run in Tealfabric. See it all in Tealfabric.”

Platform

Platform Ops and Observability — see what’s running and why

A unified operational layer across ProcessFlow, integrations, WebApps, DataPool, AI, and security—from dashboard triage to searchable logs, actionable queues, and audit trails.

Book a demoFailed sync workflow

When something breaks at 2 a.m., you start in one place and follow the thread across queues, logs, and execution history.

Distributed automation usually means distributed confusion—process failures in one tool, integration errors in another, webhook issues with no request log, data changes with no audit, and AI costs with no visibility.

Tealfabric gives operators a unified operational layer woven into the same workspace where teams design processes, connect systems, and publish apps—not a bolt-on APM product.

  • Built to be operated—not only configured; every major capability leaves a tenant-scoped trail
  • One place to follow the thread from dashboard triage to Monitor deep dive
  • Nine log types, process and integration queues, and worker throughput in one console
  • Cancel stuck jobs, re-run processes, and retry integrations—act, don't just watch
  • Execution history in Processes, Integrations, and WebApps where work was defined
  • DataPool audit and security logs in the same operational layer
  • LLM call logs with tokens, cost, latency, and 24-hour health alerts
  • Queue status API for external async polling; structured errors for alerts and retries

Security logs and governance context

Day-to-day operations

From morning health check to integration incident and audit request.

  1. dashboard alert
    step log → re-run

    Nightly sync failure

    Dashboard shows failed run → Monitor process queue → step log reveals validation error → fix and re-run.

  2. integration queue
    rate limit 429

    CRM integration degradation

    Integration queue backlog grows → worker stats confirm capacity → logs show rate-limit errors → throttle adjusted.

  3. WebApp 500s
    async stub

    Customer portal outage

    WebApp execution log shows 500s → linked process timeout → switch to async stub pattern.

  4. DataPool audit
    date range export

    Audit request

    DataPool audit export for schema changes during a date range; process and user IDs included.

  5. LLM dashboard
    p95 · tokens · cost

    AI spend control

    LLM dashboard flags high failure rate and p95 latency → LLM call logs identify problematic step.

Triage flow

Dashboard alert → Monitor logs → queue action → cross-thread trace → fix and re-run.

A nightly ERP sync fails at 2 a.m. The dashboard surfaces the failure. Monitor step logs reveal a validation error on row 847. Operators inspect the process and integration queues, trace the correlation across DataPool audit, patch the step, re-run, and confirm notification delivery.

Operations triage

Dashboard → Monitor → queue → trace → resolve

01 Dashboard02 Monitor03 Queue04 Trace05 Resolve

Home dashboard · tenant acme-ops · 06:42

Failed process runs (2)

nightly_erp_sync · VALIDATION_ERROR · 02:14

vendor_intake_submit · TIMEOUT · 04:55

Unread notifications (3)

Integration rate limit warning · hubspot-marketing

agent todos: 1 open · quick link → Monitor

Simulated flow — nightly sync failure triaged from dashboard through Monitor to fix and re-run

Capabilities

One pane for operations—act, don’t just watch.

Triage in Monitor; fix in context where the work was defined. Everything filtered by tenant—operators see their organization’s activity.

  1. failed runs
    notifications

    Operational dashboard

    Morning snapshot—notifications, failed processes, and agent todos.

    Open the platform and see what needs action. Unread alerts, recent queue failures with error snippets, and quick links to drill down in one click—not a static link farm.

  2. 9 log types
    search · sort

    Central Monitor console

    Nine searchable log types, process and integration queues, worker health.

    Process flow, step logs, integrations, connectors, DataPool audit, security, LLM calls, notification delivery, and WebApp requests—searchable, sortable, and paginated in one ops console.

  3. cancel jobs
    retry context

    Actionable queues

    View, filter, search, and cancel process and integration jobs.

    Pending, running, completed, failed, and cancelled jobs with priority, schedule time, and retry context. Cancel stuck runs when needed—act, don't just watch.

  4. Processes UI
    WebApp logs

    Feature-native history

    Execution context where work was defined—Processes, Integrations, WebApps.

    Triage in Monitor; fix in context. Process execution history grouped by run, integration test and history, WebApp per-request logs, and document review action history.

  5. LLM 24h metrics
    queue status API

    Analytics & health signals

    ProcessFlow trends, LLM ops dashboard, and queue status API.

    Tenant-scoped execution volume and failure patterns. 24-hour LLM metrics—tokens, cost, p95 latency, and automatic health alerts. External apps poll queue status without holding connections open.

  6. in-app center
    delivery log

    Notifications & alerts

    In-app center with delivery logging in Monitor.

    Unread badge, mark read, structured info/warning/success/error messages. Workflows emit notifications; operators confirm channel-level delivery success or failure.

  7. DataPool audit
    security_logs

    Governance & audit observability

    DataPool audit, security logs, and download audit on governed links.

    Which schema or record changed, by whom, from which process step. Login and security events with IP and severity. Supports compliance conversations within tenant scope.

  8. usage · billing
    org settings

    Organization usage

    Subscription and consumption visibility for tenant admins.

    Usage and billing tabs in organization settings—commercial accountability alongside technical operations.

Monitor log types

Process flow

Execution ID, status, duration, errors

Process step

Step-level log_message() output from automation code

Integrations

Integration execution queue—status, timing, failures

Connectors

Connector-level execution records

DataPool audit

Schema actions, queries, inserts/updates with context

Security

Auth and security events—IP, user, severity

LLM calls

AI request status, tokens, cost, errors

Notification delivery

Channel, recipient, delivered/failed status

WebApp

Request method, response status, duration, linked process

Operational layers

Dashboard

Triage and alerts—start here each morning

Monitor

Search, queues, workers, cross-cutting logs

ProcessFlow

Execution history, step logs, analytics, artifacts

Integrations

Queue, history, connector logs

WebApps

Request-level execution log

DataPool

Audit trail for queries and mutations

Documents

Review history

Notifications

Delivery confirmation

Org settings

Usage and billing

Fits the platform

Observability maps to how work actually flows—trigger, queue, worker execution, logs, and dashboard. ProcessFlow, integrations, WebApps, DataPool, AI agents, and notifications all leave trails you can search, sort, and act on.

Process Automation Integrations WebApps Governance

Operate with confidence

See Monitor, queues, and audit trails across your tenant.

We walk through dashboard triage, nine log types, queue cancel and re-run, DataPool audit, LLM ops metrics, and notification delivery—on the same runtime as your automation.

Book a demoStart a pilotDeveloper platform & APIs