Build Custom Stores and Destinations

30 minutes: Implement your own stores and destinations.

Prerequisites: Completed Your First Pipeline or familiar with ETL basics.

Understanding the Destination Trait

ETL delivers data to destinations in two phases:

Phase	Method	When	Data Type
Initial Copy	`write_table_rows()`	Startup	`Vec<TableRow>`
Streaming	`write_events()`	After copy	`Vec<Event>` including `Relation` schema events

Note: During initial copy, parallel table sync workers each process their own replication slot, so Begin and Commit transaction markers may appear multiple times. This does not duplicate actual row data.

Schema changes are surfaced through Event::Relation. If your destination keeps physical schemas, flush pending writes before handling a relation event, apply the supported add/drop/rename diff, then process following row events with the new schema. See Schema Changes for the current semantics and limitations.

Step 1: Create the Project

cargo new etl-custom --lib
cd etl-custom

Update Cargo.toml:

[package]
name = "etl-custom"
version = "0.1.0"
edition = "2021"

[[bin]]
name = "main"
path = "src/main.rs"

[dependencies]
etl = { git = "https://github.com/supabase/etl" }
tokio = { version = "1.0", features = ["full"] }
reqwest = { version = "0.11", features = ["json"] }
serde_json = "1.0"
tracing = "0.1"
tracing-subscriber = "0.3"

Verify: cargo check succeeds.

Step 2: Implement a Custom Store

Create src/custom_store.rs. A store must implement three traits (see Extension Points for full details):

SchemaStore - Versioned table schema storage, retrieval, and pruning
StateStore - Replication progress and destination table metadata tracking
TableLifecycleStore - Store lifecycle operations for table-copy restarts and publication changes

SharedStateStore, DestinationStore, and PipelineStore are blanket-implemented facades over these traits plus the required clone/thread-safety bounds, so custom stores do not implement those directly.

use std::collections::{BTreeMap, HashMap};
use std::sync::Arc;
use tokio::sync::Mutex;
use tracing::info;

use etl::error::EtlResult;
use etl::replication::WorkerType;
use etl::state::{
    AppliedDestinationTableMetadata, DestinationTableMetadata, TableReplicationPhase,
};
use etl::store::{SchemaStore, StateStore, TableLifecycleStore, TableReplicationStates};
use etl::store::schema::TableSchemaRetention;
use etl::types::{PgLsn, SnapshotId, TableId, TableSchema};

#[derive(Debug, Clone, Default)]
struct TableEntry {
    schemas: HashMap<SnapshotId, Arc<TableSchema>>,
    state: Option<TableReplicationPhase>,
    destination_metadata: Option<DestinationTableMetadata>,
}

#[derive(Debug, Clone)]
pub struct CustomStore {
    tables: Arc<Mutex<HashMap<TableId, TableEntry>>>,
    progress: Arc<Mutex<HashMap<WorkerType, PgLsn>>>,
}

impl CustomStore {
    pub fn new() -> Self {
        info!("creating custom store");
        Self {
            tables: Arc::new(Mutex::new(HashMap::new())),
            progress: Arc::new(Mutex::new(HashMap::new())),
        }
    }
}

impl SchemaStore for CustomStore {
    async fn get_table_schema(
        &self,
        table_id: &TableId,
        snapshot_id: SnapshotId,
    ) -> EtlResult<Option<Arc<TableSchema>>> {
        let tables = self.tables.lock().await;
        Ok(tables.get(table_id).and_then(|entry| {
            entry
                .schemas
                .iter()
                .filter(|(sid, _)| **sid <= snapshot_id)
                .max_by_key(|(sid, _)| *sid)
                .map(|(_, schema)| Arc::clone(schema))
        }))
    }

    async fn get_table_schemas(&self) -> EtlResult<Vec<Arc<TableSchema>>> {
        let tables = self.tables.lock().await;
        Ok(tables
            .values()
            .flat_map(|entry| entry.schemas.values().cloned())
            .collect())
    }

    async fn load_table_schemas(&self) -> EtlResult<usize> {
        Ok(0)
    }

    async fn store_table_schema(&self, schema: TableSchema) -> EtlResult<Arc<TableSchema>> {
        let mut tables = self.tables.lock().await;
        let id = schema.id;
        let snapshot_id = schema.snapshot_id;
        let schema = Arc::new(schema);
        tables
            .entry(id)
            .or_default()
            .schemas
            .insert(snapshot_id, Arc::clone(&schema));
        Ok(schema)
    }

    async fn prune_table_schemas(
        &self,
        table_schema_retentions: HashMap<TableId, TableSchemaRetention>,
    ) -> EtlResult<u64> {
        let mut tables = self.tables.lock().await;
        let mut removed_count = 0u64;

        for (table_id, entry) in tables.iter_mut() {
            let Some(retention) = table_schema_retentions.get(table_id) else {
                continue;
            };
            let retention_snapshot_id = SnapshotId::from(retention.to_lsn());

            let retained_snapshot_id = entry
                .schemas
                .keys()
                .filter(|snapshot_id| **snapshot_id <= retention_snapshot_id)
                .max()
                .copied();
            let Some(retained_snapshot_id) = retained_snapshot_id else {
                continue;
            };

            let before_count = entry.schemas.len();
            entry.schemas.retain(|snapshot_id, _| *snapshot_id >= retained_snapshot_id);
            removed_count =
                removed_count.saturating_add(before_count.saturating_sub(entry.schemas.len()) as u64);
        }

        Ok(removed_count)
    }
}

impl StateStore for CustomStore {
    async fn get_table_replication_state(
        &self,
        table_id: TableId,
    ) -> EtlResult<Option<TableReplicationPhase>> {
        let tables = self.tables.lock().await;
        Ok(tables.get(&table_id).and_then(|e| e.state.clone()))
    }

    async fn get_table_replication_states(
        &self,
    ) -> EtlResult<TableReplicationStates> {
        let tables = self.tables.lock().await;
        Ok(Arc::new(
            tables
                .iter()
                .filter_map(|(id, e)| e.state.clone().map(|s| (*id, s)))
                .collect::<BTreeMap<_, _>>(),
        ))
    }

    async fn load_table_replication_states(&self) -> EtlResult<usize> {
        Ok(0)
    }

    async fn update_table_replication_states(
        &self,
        updates: Vec<(TableId, TableReplicationPhase)>,
    ) -> EtlResult<()> {
        let mut tables = self.tables.lock().await;
        for (table_id, state) in updates {
            info!("table {} -> {:?}", table_id.0, state);
            tables.entry(table_id).or_default().state = Some(state);
        }
        Ok(())
    }

    async fn rollback_table_replication_state(
        &self,
        _table_id: TableId,
    ) -> EtlResult<TableReplicationPhase> {
        todo!("Implement rollback if needed")
    }

    async fn get_replication_progress(
        &self,
        worker_type: WorkerType,
    ) -> EtlResult<Option<PgLsn>> {
        let progress = self.progress.lock().await;
        Ok(progress.get(&worker_type).copied())
    }

    async fn upsert_replication_progress(
        &self,
        worker_type: WorkerType,
        flush_lsn: PgLsn,
    ) -> EtlResult<PgLsn> {
        let mut progress = self.progress.lock().await;
        let stored_lsn = progress.entry(worker_type).or_insert(flush_lsn);
        *stored_lsn = (*stored_lsn).max(flush_lsn);
        Ok(*stored_lsn)
    }

    async fn delete_replication_progress(&self, worker_type: WorkerType) -> EtlResult<()> {
        let mut progress = self.progress.lock().await;
        progress.remove(&worker_type);
        Ok(())
    }

    async fn get_destination_table_metadata(
        &self,
        table_id: TableId,
    ) -> EtlResult<Option<DestinationTableMetadata>> {
        let tables = self.tables.lock().await;
        Ok(tables
            .get(&table_id)
            .and_then(|e| e.destination_metadata.clone()))
    }

    async fn get_applied_destination_table_metadata(
        &self,
        table_id: TableId,
    ) -> EtlResult<Option<AppliedDestinationTableMetadata>> {
        self.get_destination_table_metadata(table_id)
            .await?
            .map(|metadata| metadata.into_applied())
            .transpose()
    }

    async fn load_destination_tables_metadata(&self) -> EtlResult<usize> {
        Ok(0)
    }

    async fn store_destination_table_metadata(
        &self,
        table_id: TableId,
        metadata: DestinationTableMetadata,
    ) -> EtlResult<()> {
        let mut tables = self.tables.lock().await;
        tables.entry(table_id).or_default().destination_metadata = Some(metadata);
        Ok(())
    }
}

impl TableLifecycleStore for CustomStore {
    async fn clear_table_copy_state(&self, table_id: TableId) -> EtlResult<()> {
        let mut tables = self.tables.lock().await;
        if let Some(entry) = tables.get_mut(&table_id) {
            entry.schemas.clear();
            entry.destination_metadata = None;
        }
        let mut progress = self.progress.lock().await;
        progress.remove(&WorkerType::TableSync { table_id });
        Ok(())
    }

    async fn delete_table_pipeline_state(&self, table_id: TableId) -> EtlResult<()> {
        let mut tables = self.tables.lock().await;
        tables.remove(&table_id);
        let mut progress = self.progress.lock().await;
        progress.remove(&WorkerType::TableSync { table_id });
        Ok(())
    }
}

Verify: cargo check succeeds.

Step 3: Implement a Custom Destination

Create src/http_destination.rs. A destination implements the Destination trait with four required methods:

name() - Return an identifier for logging
drop_table_for_copy() - Idempotently drop destination objects and replay state before restarting a table copy using the previously stored replicated table schema
write_table_rows() - Receive rows during initial copy together with the current replicated table schema
write_events() - Receive streaming changes (batches may span multiple tables)

There's also an optional shutdown() method with a default no-op implementation. Override it if your destination needs cleanup when the pipeline shuts down.

ETL clears its own schema versions, destination metadata, and table-sync progress only after drop_table_for_copy() succeeds. That lets the destination use the supplied previously stored replicated schema and any existing destination metadata to find the object that must be removed. If the object is already gone, return success.

use reqwest::Client;
use serde_json::json;
use std::time::Duration;
use tracing::{info, warn};

use etl::destination::{
    Destination, DropTableForCopyResult, WriteEventsResult, WriteTableRowsResult,
};
use etl::error::{ErrorKind, EtlResult};
use etl::types::{Event, ReplicatedTableSchema, TableRow};
use etl::{bail, etl_error};

#[derive(Debug, Clone)]
pub struct HttpDestination {
    client: Client,
    base_url: String,
}

impl HttpDestination {
    pub fn new(base_url: String) -> EtlResult<Self> {
        let client = Client::builder()
            .timeout(Duration::from_secs(30))
            .build()
            .map_err(|e| etl_error!(ErrorKind::Unknown, "HTTP client error", source: e))?;
        Ok(Self { client, base_url })
    }

    async fn post(&self, path: &str, body: serde_json::Value) -> EtlResult<()> {
        let url = format!("{}/{}", self.base_url.trim_end_matches('/'), path);

        for attempt in 1..=3 {
            match self.client.post(&url).json(&body).send().await {
                Ok(resp) if resp.status().is_success() => return Ok(()),
                Ok(resp) if resp.status().is_client_error() => {
                    bail!(ErrorKind::Unknown, "Client error", resp.status());
                }
                Ok(resp) => warn!("attempt {}/3: status {}", attempt, resp.status()),
                Err(e) => warn!("attempt {}/3: {}", attempt, e),
            }
            if attempt < 3 {
                tokio::time::sleep(Duration::from_millis(500 * attempt as u64)).await;
            }
        }
        bail!(ErrorKind::Unknown, "Request failed after retries");
    }
}

impl Destination for HttpDestination {
    fn name() -> &'static str {
        "http"
    }

    async fn drop_table_for_copy(
        &self,
        replicated_table_schema: &ReplicatedTableSchema,
        async_result: DropTableForCopyResult<()>,
    ) -> EtlResult<()> {
        let table_name = replicated_table_schema.name().to_string();
        info!("dropping table before copy {}", table_name);
        let result = self
            .post(&format!("tables/{table_name}/drop-for-copy"), json!({}))
            .await;
        async_result.send(result);
        Ok(())
    }

    async fn write_table_rows(
        &self,
        replicated_table_schema: &ReplicatedTableSchema,
        rows: Vec<TableRow>,
        async_result: WriteTableRowsResult<()>,
    ) -> EtlResult<()> {
        if rows.is_empty() {
            async_result.send(Ok(()));
            return Ok(());
        }
        let table_name = replicated_table_schema.name().to_string();
        info!("writing {} rows to table {}", rows.len(), table_name);

        let payload = json!({
            "table_name": table_name,
            "rows": rows.iter().map(|r| {
                json!({ "values": r.values().iter().map(|v| format!("{:?}", v)).collect::<Vec<_>>() })
            }).collect::<Vec<_>>()
        });

        let result = self.post("rows", payload).await;
        async_result.send(result);
        Ok(())
    }

    async fn write_events(
        &self,
        events: Vec<Event>,
        async_result: WriteEventsResult<()>,
    ) -> EtlResult<()> {
        if events.is_empty() {
            async_result.send(Ok(()));
            return Ok(());
        }
        info!("writing {} events", events.len());

        let payload = json!({
            "events": events.iter().map(|e| {
                match e {
                    Event::Insert(i) => json!({"type": "insert", "table": i.replicated_table_schema.name().to_string()}),
                    Event::Update(u) => json!({"type": "update", "table": u.replicated_table_schema.name().to_string()}),
                    Event::Delete(d) => json!({"type": "delete", "table": d.replicated_table_schema.name().to_string()}),
                    Event::Begin(_) => json!({"type": "begin"}),
                    Event::Commit(_) => json!({"type": "commit"}),
                    Event::Relation(r) => json!({"type": "relation", "table": r.replicated_table_schema.name().to_string()}),
                    Event::Truncate(t) => json!({"type": "truncate", "tables": t.truncated_tables.iter().map(|table| table.name().to_string()).collect::<Vec<_>>() }),
                    Event::Unsupported => json!({"type": "unsupported"}),
                }
            }).collect::<Vec<_>>()
        });

        let result = self.post("events", payload).await;
        async_result.send(result);
        Ok(())
    }
}

Verify: cargo check succeeds.

Step 4: Wire It Together

Create src/main.rs:

The custom store owns ETL runtime state. Pipeline::start() still prepares the source database with ETL's schema snapshot helpers and schema-change event trigger before replication begins.

mod custom_store;
mod http_destination;

use custom_store::CustomStore;
use etl::config::{
    BatchConfig, InvalidatedSlotBehavior, MemoryBackpressureConfig, PgConnectionConfig,
    PipelineConfig, TableSyncCopyConfig, TcpKeepaliveConfig, TlsConfig,
};
use etl::pipeline::Pipeline;
use http_destination::HttpDestination;
use std::error::Error;

#[tokio::main]
async fn main() -> Result<(), Box<dyn Error>> {
    tracing_subscriber::fmt::init();

    let pg_config = PgConnectionConfig {
        host: "localhost".to_string(),
        port: 5432,
        name: "your_database".to_string(),
        username: "postgres".to_string(),
        password: Some("your_password".to_string().into()),
        tls: TlsConfig {
            enabled: false,
            trusted_root_certs: String::new(),
        },
        keepalive: TcpKeepaliveConfig::default(),
    };

    let store = CustomStore::new();
    let destination = HttpDestination::new("https://your-endpoint.example.com".to_string())?;

    let config = PipelineConfig {
        id: 1,
        publication_name: "my_publication".to_string(),
        pg_connection: pg_config,
        batch: BatchConfig {
            max_fill_ms: 5000,
            memory_budget_ratio: 0.2,
            max_bytes: 8 * 1024 * 1024,
        },
        table_error_retry_delay_ms: 10000,
        table_error_retry_max_attempts: 5,
        max_table_sync_workers: 4,
        max_copy_connections_per_table: PipelineConfig::DEFAULT_MAX_COPY_CONNECTIONS_PER_TABLE,
        memory_refresh_interval_ms: 100,
        memory_backpressure: Some(MemoryBackpressureConfig::default()),
        table_sync_copy: TableSyncCopyConfig::default(),
        invalidated_slot_behavior: InvalidatedSlotBehavior::default(),
    };

    println!("Starting pipeline...");
    let mut pipeline = Pipeline::new(config, store, destination);
    pipeline.start().await?;
    pipeline.wait().await?;

    Ok(())
}

Note: Update the database name, password, and HTTP endpoint to match your setup.

Step 5: Test

1	`cargo run`

The pipeline will connect to Postgres and start replicating. You'll see your custom store logging state transitions and your destination receiving HTTP calls.

What You Built

Custom Store - In-memory implementation of SchemaStore, StateStore, and TableLifecycleStore
HTTP Destination - Forwards replicated data via HTTP POST with retry logic
Working Pipeline - Connects your custom components to the ETL core

Next Steps

Extension Points - Full trait API documentation
Event Types - Details on all events your destination receives
Configure Postgres - Production database setup
Architecture - How ETL works internally