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

# Call events

> Export per-call diagnostics to your organization's BigQuery table.

You need an existing BigQuery table with the schema below and a Google identity
that can read its schema and insert rows. Call event export is disabled until
you configure it in **Settings → Call events**.

1. Select **BigQuery** as the destination.
2. Enter the table as `project.dataset.table`.
3. Choose a service account and enter its email and PEM private key. On a
   self-hosted installation, you can instead select **Deployment identity** to
   use the Google credentials configured on your Dograh server.
4. Select **Test connection** to check authentication and table compatibility.
   This reads table metadata without inserting events; successful exports also
   require write permission.
5. Enable **Export call events** and select **Save**.

Each organization has its own destination and credentials. Saved private keys
are masked in the UI. Keep the masked value to retain your existing key, or
replace it to rotate the key.

**BigQuery schema**

Use these scalar columns with nullable fields; `ts` and `event` can also be
required. The adapter does not create or migrate tables.

| Column                                    | Type                         |
| ----------------------------------------- | ---------------------------- |
| `ts`                                      | TIMESTAMP                    |
| `event`                                   | STRING                       |
| `run_id`, `org_id`, `workflow_id`, `turn` | INTEGER                      |
| `severity`                                | STRING                       |
| `value_ms`                                | FLOAT                        |
| `node_id`, `node_name`                    | STRING                       |
| `detail`                                  | JSON (STRING also supported) |

You receive one row per diagnostic event. `ts` is when the event happened, and
`detail` contains JSON, stored as a native JSON value or serialized text according
to the column type. Events cover speech, accepted user turns,
mute state, tool execution, silence warnings, stage latencies, pipeline errors
and the final call summary. The transcription event records character counts
and empty/finalized flags; it does not contain transcript text or audio.

**Delivery behavior**

Events are sent after a call finishes, with their original timestamps.
Capture and export are separate from the real-time feedback logs.
Dograh stores the destination settings in its organization configuration;
event payloads are not persisted in Postgres or Redis.

Exports use bounded retries for retryable row and HTTP errors. BigQuery's
insertion IDs provide best-effort deduplication; duplicates can still occur.
Capture and pending export queues are bounded, and unsent events can be lost
on a process crash, queue overflow or exhausted retries. There is no durable
replay queue in this version. Export failures and dropped-event counts appear
in the API worker logs.

Table changes apply to new calls. An active call's batch is discarded if its
destination is changed or disabled before export, so it cannot accidentally
land in a different table. Credentials are read again before each delivery
attempt to support rotation. Requests already sent to BigQuery cannot be
recalled when you disable export.