# Service delivery report

> The 7, 30 or 90-day report: service levels, noise and correlation, remediation outcomes, compliance posture, recurrence, agent quality, and platform cost as an appendix.

_Reference · Updated 2026-10-24 · Describes release 0.1.164_

## The service delivery report

![Service delivery report over 30 days: Service levels — opened, resolved, within target and open backlog, the per-priority table, time to first analysis and to recovery, the backlog by age and the recently resolved incidents — before Noise & correlation starts; 7d / 30d / 90d and Export markdown sit in the header.](/help/media/portal/portal-report.png?v=0.1.132)

Open **Insights → Report** (`/portal/report`). Reading the report needs `report.view`; incident evidence also needs `incidents.view`, and test results need `benchmark.view`. Missing access leaves the affected source unavailable while the permitted results remain useful.

### Read it top to bottom

The **Service delivery report** answers, in order, what happened to your services in the period, how fast it was handled, what the agents changed and what still needs a person. Pick **7d**, **30d** or **90d** at the top; the choice applies to every section. Platform cost is not part of that story: it sits in a closed appendix at the end, and the month-by-month statement of the same spend is **Usage & billing**.

| Section | What it shows |
| --- | --- |
| **Service levels** | **Opened**, **Resolved**, **Within target** and **Open backlog**, a table by priority (P1 to P4, plus any unclassified incident) with opened, resolved, resolved inside target and open now, **Time to first analysis**, **Time to recovery**, the backlog by age (under a day, one to seven days, over seven) and the 25 most recently closed incidents |
| **Noise & correlation** | Alerts received against the incidents they became, **Noise reduction**, **Grouped** (incidents merged into a related incident — by the service graph, the grouping agent or a person — counted with it, not as incidents), **Suppressed** and **Ignored** alerts |
| **Remediation outcomes** | Changes the platform proposed or made: pre-approved automatic against human-approved, **Validated**, **Completed, not verified**, **Pending approval**, **Failed** and **Rolled back**, by risk tier, with the mean **Approval turnaround** |
| **Compliance posture** | The scanner-cleared score now against the first scored day of the period, unresolved findings by severity, **Suppressed** findings and findings **Fixed — awaiting CSPM**, with a link to **Security posture** |
| **Recurrence & problems** | Open problems, automation candidates, problems **Blocked** because a recorded fix did not hold, and the most recurring |
| **Human feedback & agent quality** | **Analysed automatically**, **Root cause written**, **Test Lab pass rate** (the lowest pillar, each pillar's newest finished run), **Reviewer score** and the latest Test Lab result per pillar |
| **Appendix · Platform cost (FinOps)** | Closed by default: cost per analysis, the last 30 days of spend per agent layer with its pricing caveats, and the costliest incidents. Test runs are excluded. AI spend is a platform cost, not a service outcome; the same run records by calendar month are on **Usage & billing** |

Compliance findings are not counted as incidents (they are the Security posture worklist), and neither are incidents a person closed as noise. A resolved incident is within target when its condition cleared inside the target for its priority: P1 4 hours, P2 8, P3 24, P4 72 unless your workspace sets `sla_targets_h`; an unclassified incident has no target. Time to recovery uses the close time only for an incident without a verified clear or a source resolution. Approval turnaround pairs each approval request with its first decision. Counts come from one page of up to 500 rows per list: a full page shows a **+** after the figure, because the period may hold more. A section whose data your role may not read says which permission it needs and the rest of the report stays.

**Export markdown** downloads the same sections in the same order, platform cost last, with the pricing caveats. It is a local download and does not create an artifact.

## Use cases

### A monthly service report — executive

Open **Insights → Report** and read it top to bottom: service levels, noise and correlation, remediation outcomes, compliance posture, recurrence and problems, then human feedback and agent quality. A **+** after a count means the period may hold more than one page of rows. Open **Appendix · Platform cost (FinOps)** only when cost is the question — or **Usage & billing** when the question is a calendar month's statement — then use **Export markdown** to download the same document.

## Troubleshooting

Symptoms specific to this page. Sign-in problems, missing nav entries in general, stuck or queued investigations, and `$0.0000` costs are in the **Troubleshooting** chapter.

| Symptom | Cause | What to do |
| --- | --- | --- |
| **Latest benchmarks** missing from the report | No `benchmark.view`, or no graded run yet | Grant the permission, or run the builtin suite once. |

## Next steps

- [Usage and billing](/help/usage-and-billing.md) — the same spend by calendar month.
- [Feedback](/help/feedback.md) — the reviewer scores the report summarises.
- [Benchmarks and scores](/help/benchmarks-and-scores.md) — where the Test Lab pass rate comes from.
- [Knowledge, x-ray and insights](/help/intelligence.md) — the hub for every page of this group.