Struct CallBatchLayer

pub struct CallBatchLayer { /* private fields */ }

Provider layer that aggregates contract calls (eth_call) over a time period into a single Multicall3 contract call.

Some methods, such as eth_getBlockNumber, are first converted into contract calls to the Multicall3 contract itself and then aggregated with other eth_calls.

Only calls that:

• target the latest block ID,
• have no state overrides,
• have a target address and calldata,
• have no other properties (nonce, gas, etc.)

can be sent with a multicall. This of course requires that the Multicall3 contract is deployed on the network, by default at MULTICALL3_ADDRESS.

This layer is useful for reducing the number of network requests made. However, this only works when requests are made in parallel, for example when using the tokio::join! macro or in multiple threads/tasks, as otherwise the requests will be sent one by one as normal, but with an added delay.

Examples

use alloy_provider::{layers::CallBatchLayer, Provider, ProviderBuilder};
use std::time::Duration;

// Build a provider with the default call batching configuration.
let provider = ProviderBuilder::new().with_call_batching().connect(url).await?;

// Build a provider with a custom call batching configuration.
let provider = ProviderBuilder::new()
    .layer(CallBatchLayer::new().wait(Duration::from_millis(10)))
    .connect(url)
    .await?;

// Both of these requests will be batched together and only 1 network request will be made.
let (block_number_result, chain_id_result) =
    tokio::join!(provider.get_block_number(), provider.get_chain_id());
let block_number = block_number_result?;
let chain_id = chain_id_result?;
println!("block number: {block_number}, chain id: {chain_id}");

Implementations

impl CallBatchLayer

pub const fn new() -> Self

Create a new CallBatchLayer with a default wait of 1ms.

pub const fn wait(self, wait: Duration) -> Self

Set the amount of time to wait before sending the batch.

This is the amount of time to wait after the first request is received before sending all the requests received in that time period.

This means that every request has a maximum delay of wait before being sent.

The default is 1ms.

pub const fn multicall3_address(self, m3a: Address) -> Self

Set the multicall3 address.

The default is MULTICALL3_ADDRESS.

pub const fn arbitrum_compat(self) -> Self

Use the Arbitrum ArbSys precompile for block number queries.

On Arbitrum, block.number returns the parent chain’s block number (L1). Without this setting, batched eth_blockNumber calls through Multicall3 will therefore return the wrong value. Enabling this queries the L2 block number via ArbSys instead.

The default is false. This should only be enabled when interacting with Arbitrum rollups.

Trait Implementations

impl Debug for CallBatchLayer

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for CallBatchLayer

fn default() -> Self

Returns the “default value” for a type.

impl<P, N> ProviderLayer<P, N> for CallBatchLayer

where P: Provider<N> + 'static, N: Network,

type Provider = CallBatchProvider<P, N>

The provider constructed by this layer.

fn layer(&self, inner: P) -> Self::Provider

Wrap the given provider in the layer’s provider.