Skip to content
Just for: Scrum Masters Product Owners Program Managers Project Managers Product Managers Everyone else

Report · Incident Management

Incident trends

Incidents per week by severity and the trend in mean time to restore, July to September 2026. Demo data.

10 weeks Without a Sev 1 incident

Demo data

Fewer incidents, and faster recovery

Weeks of Jul 6 to Sep 28, 2026. As of October 4, 2026. Source: Pecos Mutual incident tool export (fictional).

Summary

Incidents fell steadily through the quarter. Weekly incidents averaged 16 in the first four weeks and 9.5 in the last four. We had one Sev 1, in the week of Jul 20, and none in the 10 weeks since. When Sev 1 and Sev 2 incidents do happen, we now restore service much faster.

Charts

  • Sev 1
  • Sev 2
  • Sev 3
  • Sev 4
Incidents per week by severityWeekly incidents fell from an average of 16 in the first four weeks to 9.5 in the last four. The only Sev 1 was in the week of Jul 20.05101520Jul 6, Sev 2: 2Jul 6, Sev 3: 6Jul 6, Sev 4: 9Jul 6Jul 13, Sev 2: 1Jul 13, Sev 3: 5Jul 13, Sev 4: 8Jul 13Jul 20, Sev 1: 1Jul 20, Sev 2: 2Jul 20, Sev 3: 7Jul 20, Sev 4: 10Jul 20Jul 27, Sev 2: 1Jul 27, Sev 3: 5Jul 27, Sev 4: 7Jul 27Aug 3, Sev 2: 1Aug 3, Sev 3: 6Aug 3, Sev 4: 8Aug 3Aug 10, Sev 2: 2Aug 10, Sev 3: 4Aug 10, Sev 4: 7Aug 10Aug 17, Sev 2: 1Aug 17, Sev 3: 5Aug 17, Sev 4: 6Aug 17Aug 24, Sev 3: 4Aug 24, Sev 4: 7Aug 24Aug 31, Sev 2: 1Aug 31, Sev 3: 4Aug 31, Sev 4: 6Aug 31Sep 7, Sev 2: 1Sep 7, Sev 3: 3Sep 7, Sev 4: 6Sep 7Sep 14, Sev 2: 2Sep 14, Sev 3: 4Sep 14, Sev 4: 5Sep 14Sep 21, Sev 2: 2Sep 21, Sev 3: 3Sep 21, Sev 4: 5Sep 21Sep 28, Sev 3: 3Sep 28, Sev 4: 4Sep 28
Weekly incidents fell from an average of 16 in the first four weeks to 9.5 in the last four. The only Sev 1 was in the week of Jul 20. Demo data, Weeks of Jul 6 to Sep 28, 2026.
Mean time to restore, Sev 1–2 incidents (hours)Mean hours to restore service for Sev 1 and Sev 2 incidents fell from 4.5 in the week of Jul 6 to 1.4 in the week of Sep 21. Weeks with no Sev 1–2 incidents have no value.0h2h4h6hJul 6Jul 20Aug 3Aug 17Aug 31Sep 14Sep 28MTTR, Jul 6: 4.5hMTTR, Jul 13: 3.8hMTTR, Jul 20: 5.2hMTTR, Jul 27: 3.1hMTTR, Aug 3: 2.9hMTTR, Aug 10: 2.4hMTTR, Aug 17: 2hMTTR, Aug 31: 0.3hMTTR, Sep 7: 0.7hMTTR, Sep 14: 0.5hMTTR, Sep 21: 1.4hMTTR 1.4h
Mean hours to restore service for Sev 1 and Sev 2 incidents fell from 4.5 in the week of Jul 6 to 1.4 in the week of Sep 21. Weeks with no Sev 1–2 incidents have no value. Demo data, Weeks of Jul 6 to Sep 28, 2026.

Findings

  • Fewer low-severity incidents. Sev 3 and Sev 4 incidents account for most of the fall. Several of them came from alerts that post-incident reviews have since retuned.
  • Faster recovery. Mean time to restore for Sev 1–2 incidents fell from 4.5 hours in the week of Jul 6 to under 1.5 hours in every week of September.
  • The Sep 21 week was the exception. A 150-minute Underwriting Workbench outage pulled that week's mean up to 1.4 hours. Its post-incident review found a missed certificate renewal.
  • Six Sev 2 incidents in September, across Claims Intake, Payments, Agent Portal and Underwriting Workbench. They match the downtime in the Platform Health scorecard.

Actions

  • Scrum Masters: keep running post-incident reviews within 5 working days. Their alert changes are behind most of the drop in Sev 3 and Sev 4 incidents.
  • Underwriting Workbench Home Team: automate certificate renewal and alert 30 days before expiry. Owner: Policy Platform Product Owner.
  • Program Manager: report open review actions monthly at the Platform Health review until they are closed.

Data

Incidents per week by severity, and mean time to restore (demo data, as of 2026-10-04)
Week ofSev 1Sev 2Sev 3Sev 4TotalMTTR Sev 1–2 (hours)
Jul 60269174.5
Jul 130158143.8
Jul 2012710205.2
Jul 270157133.1
Aug 30168152.9
Aug 100247132.4
Aug 170156122
Aug 24004711–
Aug 310146110.3
Sep 70136100.7
Sep 140245110.5
Sep 210235101.4
Sep 2800347–

Source: Pecos Mutual incident tool export (fictional). As of October 4, 2026. All figures are fictional demo data.

Keep going

Related pages

Playbook

Triage and declare an incident

How to check a problem's impact, set its severity, declare the incident and get the right people working on it within minutes.

Playbook

Run a post-incident review

How to hold a blameless review after a Sev 1 or Sev 2 incident and turn what we learned into tracked actions.

Report

Platform Health scorecard

Availability against target and error budget burn for six platforms in September 2026. Demo data.