IR node starts longer than expected

In AIO IR node starts in seconds, while in neofs-testcases it takes a minute to start:

COMMAND: ./neofs-cli --config /Users/evgeniiz/Projects/neofs-testcases-nspcc/neofs-testcases/test-run-2025-11-11-23-48-32-205563/env_files/neofs-env-2025-11-11-23-48-33-8454031360/inner_ring_bhritstcea/ir_cli_config_tzngzjeelc.yml control healthcheck --endpoint 'localhost:49700' --ir
RETCODE: 1

STDOUT:
Error: failed to dial "localhost:49700" via gRPC: connection error: desc = "transport: error while dialing: dial tcp [::1]:49700: connect: connection refused"

STDERR:

Start / End / Elapsed	 23:48:46.963606 / 23:48:47.515637 / 0:00:00.552031

....

COMMAND: ./neofs-cli --config /Users/evgeniiz/Projects/neofs-testcases-nspcc/neofs-testcases/test-run-2025-11-11-23-48-32-205563/env_files/neofs-env-2025-11-11-23-48-33-8454031360/inner_ring_bhritstcea/ir_cli_config_tzngzjeelc.yml control healthcheck --endpoint 'localhost:49700' --ir
RETCODE: 1

STDOUT:
Health status: INITIALIZING_NETWORK

STDERR:

Start / End / Elapsed	 23:49:39.165310 / 23:49:39.425300 / 0:00:00.259990

Timestamps: "23:48:46.963606" -> "23:49:39.165310"

IR logs:
ir_stderr.txt
IR config:
ir_node_config_vrqilkkrlc.yml

Need to investigate the delay since it significantly slows the tests down.

Reproduce the delay with the neofs-cli control healthcheck command and compare the reported timestamps. Inspect the attached IR logs and ir_node_config_vrqilkkrlc.yml, then trace the startup path responsible for INITIALIZING_NETWORK. Done means the cause of the roughly one-minute delay is identified and the startup behavior is improved or documented with evidence.