> ## Documentation Index
> Fetch the complete documentation index at: https://86d.store/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Watch a Store Runtime

> Five different things get called analytics. Here is who owns each one, where its data goes, and which of them you can trust for revenue.

<Warning>
  **In development.** 86d is being built in the open. Every capability is Experimental until it earns evidence, so check [maturity levels](/docs/resources/versioning) before you rely on anything here.
</Warning>

The word "analytics" covers at least five unrelated things, owned by different people and sent to different places. Conflating them is how a merchant ends up reconciling their books against a browser event. 86d names each one.

## Five kinds of data

| Kind                            | Who owns it, and why                                                | Where it goes                                              |
| ------------------------------- | ------------------------------------------------------------------- | ---------------------------------------------------------- |
| **Storefront Analytics**        | You. Shopper behavior and commerce reporting for one Store          | Your Store's database, or somewhere you chose              |
| **Third-party Analytics**       | You. Selected Storefront events exported outward                    | Services you configured, such as Google Tag Manager or GA4 |
| **Store Runtime diagnostics**   | You. Errors, logs, traces, health, performance for one deployment   | Your logs, or an error project you own                     |
| **Managed Runtime Diagnostics** | 86d, only when you explicitly turn it on. Bounded errors and health | A disclosed 86d operational project                        |
| **86d Observability**           | 86d. The health of 86d.app itself                                   | 86d operational systems                                    |

None of it is commerce truth. A browser event never creates or repairs an [Order](/docs/modules/orders), a [Payment](/docs/modules/payments), an [Inventory](/docs/modules/inventory) count, a [Loyalty](/docs/modules/loyalty) balance, revenue, or GMV. When a dashboard and the database disagree, the database is right.

## Logs

The [Store Runtime](/docs/concepts/architecture) writes Next.js and application logs to standard output.

* In Docker Compose, use `docker compose logs -f store`.
* On Vercel, use the project's log view or `vercel logs`.
* On Railway, use the service log view.

There is no log shipper in the box. If you host it yourself, use your provider's log forwarding.

## Error reporting

Set a DSN and unhandled errors reach a Sentry project you control:

```bash .env theme={null}
SENTRY_DSN=https://your_sentry_key_here@your_org_here.ingest.sentry.io/your_project_id_here
NEXT_PUBLIC_SENTRY_DSN=https://your_sentry_key_here@your_org_here.ingest.sentry.io/your_project_id_here
```

The integration initializes only when the DSN exists, so leaving these unset costs you nothing. Sampling and data handling are yours to decide.

<Warning>
  `NEXT_PUBLIC_SENTRY_DSN` is public by design, and the events it sends are not. Turn on Sentry scrubbing, and keep secrets, payment credentials, and Shopper personal data out of your error context and breadcrumbs.
</Warning>

Managed Runtime Diagnostics is a separate thing, off by default, and enabled only by the exact value `managed-runtime-diagnostics-v1` in `86D_TELEMETRY`. It sends a bounded health and error contract to a disclosed 86d endpoint. It does not send Shopper behavior, commerce events, or your sales figures.

## Sending events to your own tools

[Third-party Analytics](/docs/resources/glossary#third-party-analytics) pushes selected [Storefront](/docs/concepts/storefront) events to services you picked. Set your Google Tag Manager container:

```bash .env theme={null}
NEXT_PUBLIC_GOOGLE_TAG_MANAGER_ID=GTM-0000000
```

The Storefront pushes browser events to `dataLayer` when the container exists. Whether Tag Manager forwards them to GA4 or anywhere else is your configuration, not 86d's.

Event names include `view_item`, `add_to_cart`, `begin_checkout`, and `purchase`. Every one of them is a report, not a record. A `purchase` event in the browser proves a script ran. It does not prove a Payment settled or an Order exists.

## The Analytics Module

[`@86d-store/analytics`](/docs/modules/analytics) records Storefront Analytics events in your Store's database and renders funnel views for you.

```bash theme={null}
86d module add analytics
```

It reads. It does not write commerce facts and it cannot repair one.

## Audit records

[`@86d-store/audit-log`](/docs/modules/audit-log) holds administrative records for one Store Runtime.

An append-oriented table is not a tamper-evident audit trail, and it is worth being precise about that with anyone who asks. A Stable audit record needs a server-authored actor, scope, target, action, outcome, and approval provenance, with access controls that have been tested. That is not where this Module is today.

## Health endpoint

Every Store Runtime serves one:

```text theme={null}
GET /api/health
```

Use it for deployment health checks. It tells you the process is up. It does not tell you [Checkout](/docs/modules/checkout) works, or that a provider [Integration](/docs/concepts/connections) is reachable.

## Local diagnosis

`86d doctor` reports pass, warn, or fail for runtime prerequisites, environment variables, database connectivity, generated files, and TypeScript configuration.

```bash theme={null}
86d doctor
```

The CLI collects no usage data. If that ever changes it will arrive with its own name, a published data contract, a named destination, and a way to turn it off.

## Related pages

* [Analytics](/docs/modules/analytics)
* [Audit log](/docs/modules/audit-log)
* [Secure a Store Runtime](/docs/operations/security)
* [How commerce records relate](/docs/concepts/commerce-model)
* [How 86d separates product authority](/docs/concepts/architecture)
* [Glossary](/docs/resources/glossary#storefront-analytics)