Nautilus with Bybit adapter silently dies if bars are requested for long period of time

[h3x3d]

Bug Report

Confirmed

• I’ve re-read the relevant sections of the documentation.
• I’ve searched existing issues and discussions to avoid duplicates.
• I’ve reviewed or skimmed the source code (or examples) to confirm the behavior is not by design.
• I’ve tested this issue using a recent development wheel (dev develop or a nightly) and can still reproduce it.

Expected Behavior

I was expecting to succesfully request bars (m1, m5) from bybit exchange for few days.

Actual Behavior

Application silently fails if i request m5 or m1 bars for a long period of time.

Steps to Reproduce the Problem

1. Run provided strategy with provided runner
2. Process just silently fails for some reason after receiving bar response

Code Snippets or Logs

Strategy:

class RandomConfig(StrategyConfig, frozen=True):
    instrument_id: InstrumentId


class RandomStrategy(Strategy):
    order = None
    counter = 0
    instrument: Instrument
    bar_type: BarType
    window: int = 5
    orders: dict = {}

    def __init__(self, config: RandomConfig) -> None:
        super().__init__(config)

        self.bar_type = BarType(
            instrument_id=self.config.instrument_id,
            bar_spec=BarSpecification(
                step=5,
                aggregation=BarAggregation.MINUTE,
                price_type=PriceType.LAST,
            ),
            aggregation_source=AggregationSource.EXTERNAL
        )

    def on_start(self) -> None:
        self.instrument = self.cache.instrument(self.config.instrument_id)

        self.request_bars(
            bar_type=self.bar_type,
            start=self.clock.utc_now() - pd.Timedelta(days=15)
        )

Runner:

    instrument_provider_config = InstrumentProviderConfig(load_all=True)

    data_clients = {
        BYBIT: BybitDataClientConfig(
            testnet=False,
            instrument_provider=instrument_provider_config,
            product_types=[BybitProductType.LINEAR]
        )
    }
    node_config = TradingNodeConfig(
        trader_id="STRATEGY-000001",
        exec_clients = {
            BYBIT: SandboxExecutionClientConfig(
                venue=BYBIT,
                starting_balances=["100 USDT"],
                instrument_provider=instrument_provider_config,
                account_type="MARGIN",
                oms_type="HEDGING",
            )
        },

        data_clients = data_clients,

        timeout_connection=30.0,
        timeout_reconciliation=10.0,
        timeout_portfolio=10.0,
        timeout_disconnection=10.0,
        timeout_post_stop=5.0,
    )

    node = TradingNode(config=node_config)

    node.add_data_client_factory(BYBIT, BybitLiveDataClientFactory)
    node.add_data_client_factory(BINANCE, BinanceLiveDataClientFactory)

    for exec_client in node_config.exec_clients:
        node.add_exec_client_factory(exec_client, SandboxLiveExecClientFactory)

    node.trader.add_strategy(RandomStrategy(RandomConfig(
        instrument_id=InstrumentId.from_str("XPINUSDT-LINEAR.BYBIT")
    )))

    node.build()
    node.run()

Log (tail only):

2025-10-28T07:15:22.930670603Z [INFO] STRATEGY-000001.RandomStrategy: RUNNING
2025-10-28T07:15:22.930679727Z [INFO] STRATEGY-000001.STRATEGY-000001: RUNNING
2025-10-28T07:15:22.930685314Z [INFO] STRATEGY-000001.TradingNode: RUNNING
2025-10-28T07:15:22.930802504Z [INFO] STRATEGY-000001.DataClient-BYBIT: Request XPINUSDT-LINEAR.BYBIT-5-MINUTE-LAST-EXTERNAL bars from 2025-10-13T07:15:22.930539000Z to 2025-10-28T07:15:22.930572164Z
2025-10-28T07:15:25.381041233Z [INFO] STRATEGY-000001.RandomStrategy: Received <Bar[4320]> data for XPINUSDT-LINEAR.BYBIT-5-MINUTE-LAST-EXTERNAL
(.nix-venv) 
[xxx@nuc:~/source/trading/labs]$ 

Specifications

• OS platform: Linux
• Python version: 3.13.7
• nautilus_trader version: 1.221.0a20251024

PS:

Looks like this happens when pagination is involved (number of bars is more then 1000). Sorting issue.

[cjdsellers]

Hi @h3x3d

Thanks for the report.

The Bybit adapter was actually ported to Rust a couple of days ago on develop branch (post your report). It's possible the behavior has changed now and this is no longer an issue?