PagerDuty + UAIO

Stop Getting Paged at 3 AM.

iTechSmart UAIO intercepts PagerDuty incidents before they escalate. The problem is classified, diagnosed, and resolved autonomously — with proof-of-fix — while your engineers sleep through the night.

See How It WorksView Code
0.6h
Median Time to Resolve
90%+
Auto-Resolve Rate
0
Engineers Woken Up
100%
Proof-of-Fix Attached

Four Steps to Zero Pages

01

PagerDuty Incident Created

An incident fires in PagerDuty from any source: Datadog, CloudWatch, Prometheus, custom monitors. The webhook extension sends the payload to UAIO before escalation starts.

02

Intercept Before Escalation

UAIO receives the incident within seconds. Classification begins immediately. Your escalation policy pauses while autonomous remediation runs, so no one gets paged yet.

03

Autonomous Remediation

The incident is classified, root cause is identified, and the matching runbook executes. UAIO captures terminal output, log diffs, and before/after system state throughout.

04

Auto-Resolve with Proof

Once the fix is verified, UAIO resolves the PagerDuty incident via the API with a full proof-of-fix note. Your on-call engineer never knew it happened.

Incident
Intercept
Remediate
Auto-Resolve

Drop-In Webhook Handler

Deploy this handler as an Express route or serverless function. It classifies via the iTechSmart API and auto-resolves the PagerDuty incident when confidence is high.

JavaScript · Webhook Handler
// PagerDuty Webhook Handler — iTechSmart UAIO
// Deploy as a serverless function or Express route

const ITECHSMART_API = "https://api.itechsmart.dev/v1/classify";
const ITECHSMART_KEY = process.env.ITECHSMART_API_KEY;
const PAGERDUTY_KEY  = process.env.PAGERDUTY_API_KEY;

export async function handlePagerDutyWebhook(req, res) {
  const event = req.body.event;

  // Only act on new incidents
  if (event.event_type !== "incident.triggered") {
    return res.status(200).json({ skipped: true });
  }

  const incident = event.data;

  // Step 1: Send to iTechSmart UAIO for classification
  const classify = await fetch(ITECHSMART_API, {
    method: "POST",
    headers: {
      "Authorization": `Bearer ${ITECHSMART_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      source: "pagerduty",
      incident_id: incident.id,
      title: incident.title,
      urgency: incident.urgency,
      service: incident.service?.summary,
      description: incident.body?.details || incident.title,
      escalation_policy: incident.escalation_policy?.summary,
      created_at: incident.created_at,
    }),
  });

  const result = await classify.json();

  // Step 2: If confidence is high enough, auto-resolve
  if (result.confidence >= 0.90 && result.remediation_status === "success") {
    await fetch(`https://api.pagerduty.com/incidents/${incident.id}`, {
      method: "PUT",
      headers: {
        "Authorization": `Token token=${PAGERDUTY_KEY}`,
        "Content-Type": "application/json",
        "From": "[email protected]",
      },
      body: JSON.stringify({
        incident: {
          type: "incident_reference",
          status: "resolved",
          resolution: [
            "[iTechSmart UAIO] Auto-resolved",
            `Classification: ${result.classification}`,
            `Confidence: ${(result.confidence * 100).toFixed(1)}%`,
            `Root Cause: ${result.root_cause}`,
            `Proof: ${result.proof_of_fix}`,
          ].join("\n"),
        },
      }),
    });

    return res.status(200).json({
      resolved: true,
      classification: result.classification,
      proof: result.proof_of_fix,
    });
  }

  // Confidence too low — let escalation proceed normally
  return res.status(200).json({
    resolved: false,
    reason: "below_confidence_threshold",
    confidence: result.confidence,
  });
}

Frequently Asked Questions

Does this replace PagerDuty?

No. PagerDuty remains your incident management and escalation platform. iTechSmart UAIO sits between your monitors and your engineers: it intercepts incidents, fixes them, and resolves them via the PagerDuty API. If UAIO cannot fix an incident, escalation proceeds normally and your on-call engineer gets paged as usual.

What happens if remediation fails?

UAIO never suppresses a real incident. If remediation fails or confidence is below your threshold, the PagerDuty escalation policy resumes immediately. The incident gets a note with full diagnostic context: what was attempted, what failed, and recommended manual steps. Your engineer gets paged with more context than they would have had otherwise.

Can my team see what UAIO did?

Yes. Every auto-resolved incident includes a detailed proof-of-fix note directly in PagerDuty: classification, root cause, remediation steps taken, terminal output, log diffs, and before/after system state. Your team can review every action UAIO took in the PagerDuty timeline. Nothing is hidden.

Is this safe for production?

Yes. UAIO runs with scoped permissions you define. You control which services, urgency levels, and escalation policies UAIO can intercept. Most customers start with low-urgency incidents on non-critical services, validate the results, and expand from there. UAIO also has a dry-run mode that classifies without remediating.

Ready to Let Your Team Sleep?

Connect PagerDuty to iTechSmart UAIO and stop waking up engineers for problems that fix themselves. Start with low-urgency incidents and expand when you trust it.

Book a DemoAPI Docs