lssa/tools/e2e_bench/README.md

# e2e_bench

End-to-end LEZ scenarios driven through the wallet against an in-process sequencer + indexer wired to an external Bedrock node. Times each step (submit, inclusion, wallet sync) and records borsh sizes for every block produced, split into per-tx-variant counts.

## Run

Required env vars (no defaults):

```sh
export LEZ_BEDROCK_BIN=/path/to/logos-blockchain/target/release/logos-blockchain-node
export LEZ_BEDROCK_CONFIG_DIR=/path/to/bedrock/configs
# optional: LEZ_BEDROCK_PORT (default 18080)
```

The config dir must contain `node-config.yaml` and a `deployment-settings.yaml` template with the literal string `PLACEHOLDER_CHAIN_START_TIME` (rewritten per launch).

```sh
# All scenarios, dev-mode proving (fast)
RISC0_DEV_MODE=1 cargo run --release -p e2e_bench -- --scenario all

# One scenario, real proving (slow)
cargo run --release -p e2e_bench -- --scenario amm
```

Scenarios: `token`, `amm`, `fanout`, `private`, `parallel`, `all`.

## What you'll see

Per scenario: a step table (`submit_s`, `inclusion_s`, `sync_s`, `total_s`) and a size summary covering every block captured during the scenario (block_bytes total/mean/min/max; per-tx-variant sizes for public, PPE, and program-deployment transactions).

The fanout, parallel, and private scenarios are the most representative for L1-payload-size measurements since they put multiple txs per block.

JSON output is written to `target/e2e_bench.json`.
feat: add e2e_bench tool for end-to-end scenario latency, block, and tx-byte measurements 2026-05-19 00:36:31 +02:00			`# e2e_bench`

			`End-to-end LEZ scenarios driven through the wallet against an in-process sequencer + indexer wired to an external Bedrock node. Times each step (submit, inclusion, wallet sync) and records borsh sizes for every block produced, split into per-tx-variant counts.`

			`## Run`

			`Required env vars (no defaults):`

			```sh
			`export LEZ_BEDROCK_BIN=/path/to/logos-blockchain/target/release/logos-blockchain-node`
			`export LEZ_BEDROCK_CONFIG_DIR=/path/to/bedrock/configs`
			`# optional: LEZ_BEDROCK_PORT (default 18080)`
			```

			The config dir must contain `node-config.yaml` and a `deployment-settings.yaml` template with the literal string `PLACEHOLDER_CHAIN_START_TIME` (rewritten per launch).

			```sh
			`# All scenarios, dev-mode proving (fast)`
			`RISC0_DEV_MODE=1 cargo run --release -p e2e_bench -- --scenario all`

			`# One scenario, real proving (slow)`
			`cargo run --release -p e2e_bench -- --scenario amm`
			```

			Scenarios: `token`, `amm`, `fanout`, `private`, `parallel`, `all`.

			`## What you'll see`

refactor(e2e_bench)!: Duration-typed timings, seconds-float JSON, tokio::timeout BREAKING CHANGE: bench JSON renames per-step / per-scenario timing fields from _ms (float milliseconds) to _s (float seconds). Renames: submit_ms → submit_s, inclusion_ms → inclusion_s, wallet_sync_ms → wallet_sync_s, total_ms → total_s, setup_ms → setup_s, bedrock_finality_ms → bedrock_finality_s, total_wall_seconds → total_wall_s. measure_bedrock_finality timeout floor also shifts slightly: on timeout the field is now ~60.000s rather than "first poll tick past 60s". 2026-05-19 22:59:02 +02:00			Per scenario: a step table (`submit_s`, `inclusion_s`, `sync_s`, `total_s`) and a size summary covering every block captured during the scenario (block_bytes total/mean/min/max; per-tx-variant sizes for public, PPE, and program-deployment transactions).
feat: add e2e_bench tool for end-to-end scenario latency, block, and tx-byte measurements 2026-05-19 00:36:31 +02:00
			`The fanout, parallel, and private scenarios are the most representative for L1-payload-size measurements since they put multiple txs per block.`

			JSON output is written to `target/e2e_bench.json`.