Send2Bucket
Send2Bucket is a hosted ingestion layer. You send JSON events or OpenTelemetry logs, traces and metrics to https://api.s2b.tznc.dev, and Send2Bucket writes them to a bucket or table in your own account.
How it works
- Your client sends a request. Send2Bucket stores the batch durably and only then answers
200. - Send2Bucket groups batches per destination and delivers them at 16 MiB of uncompressed data or every 60 seconds, whichever comes first.
- It writes gzip-compressed NDJSON to your bucket, or appends rows to your table through its catalog, and records the delivery against each batch.
Terms
| Term | Meaning |
|---|---|
| Organization | The billing unit. Usage from all its streams is added together. |
| Ingestion stream | One sending project's endpoint: an API key whose accepted data is delivered to exactly one destination. Called a stream for short. It is not a message stream: there are no consumers, offsets or replay; your bucket or table is where data is read. |
| API key | s2b_ingest_ followed by 32 base62 characters. The dashboard shows only its first 14 characters after creation. |
| Destination | The bucket or table a stream writes to. See Destinations. |
| Batch | The records accepted for one stream from one request, identified by batch_id (bat_…). The batch format creates one batch per stream. |
| Delivery batch | One object or one table commit, identified by delivery_batch_id (dlv_…). It can hold records from several batches. |
What you own
You own the bucket or catalog, its region, retention and access policies, and you pay its storage and query fees. Send2Bucket holds payloads only until they are delivered. Data already delivered stays in your account after you disconnect the stream.
Next
- Quick start: create an ingestion stream and send the first event.
- API reference: endpoints, errors and limits.
- Delivery format: object keys, the envelope and table columns.
- Querying: read your data with DuckDB, Trino, BigQuery, PostgreSQL or MySQL.