> ## Documentation Index
> Fetch the complete documentation index at: https://docs.aciona.me/llms.txt
> Use this file to discover all available pages before exploring further.
# Operational report
> How to read MTTA, MTTR, the 5-minute SLA, the escalation rate and volume — and how to compare two periods.
The **Operational report** lives under **Operational → Report** in the dashboard. It answers a single question: **are we getting better?**
Besides the figures for the selected period, the page fetches a **second window** and shows how each indicator moved. That comparison is what turns "our time to acknowledge is 3m20s" into "our time to acknowledge dropped 41% since week 0".
## The five indicators
| Indicator | What it is | Better when |
| - | - | - |
| **Time to acknowledge (MTTA)** | Median time between the first alert received and the moment someone acknowledged | lower |
| **Time to resolve (MTTR)** | Median time between the first alert received and the resolution | lower |
| **Acknowledged within 5 minutes** | Share of **critical** incidents acknowledged within 5 minutes | higher |
| **Escalation rate** | Share of incidents that had to move up a level | lower |
| **Incident volume** | Total for the period | **neither better nor worse** |
Volume is deliberately left uncoloured: more incidents can mean broader coverage (good) or more instability (bad), and the report has no way to tell which.
## Slices
* **Period**: last 7, 14, 30 (default) or 90 days, or a custom window of up to 180 days.
* **Team**: all teams, a specific team, or **No team** — needed to see the incidents with no team attached.
* **Compare with**: **Previous period** (same length, ending where the current period starts), **Custom period** (for example week 0 of a pilot) or **Do not compare**.
Dates and times are shown in the timezone configured in your account.
## Reading the change
* A duration change is shown in seconds and percent: `−80s · −40%`.
* A change of a percentage is shown in **percentage points** (`+16.7 pp`), never as a "percent of a percent".
* Every change carries an **arrow** and an accessible description on top of the colour — it stays readable in greyscale and with a screen reader.
When either side has no sample, the change shows as `—` with no arrow. The report never treats "not measured" as zero: that would invent an improvement that did not happen.
## How we read these numbers
The **"How we read these numbers"** block, right below the cards, shows what each metric leaves out with the exact count:
* Test incidents, ignored
* Incidents without a linked alert
* Never acknowledged (outside MTTA)
* Still open (outside MTTR)
* Auto-resolved by the source, with its share and its own median
The same incident can appear on more than one of those lines — they do **not** add up to the period total.
## Volume by service
The table lists incidents, alerts and the medians per service, in the order the product returns (most incidents first). You can reorder by clicking a header. The **No service** row always stays last, and the footer reconciles with the period total — that is how you check the numbers by hand.
## Export
The **Export CSV** button downloads the **current** period with the same period and team filters. The comparison window is not exported.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.