Implement some much-needed improvements for the UHS load-generators

[HalosGhost]

This first commit brings targeted-load for the 2PC load-gen; another commit will follow updating the atomizer load-gen as well.

cf. #279

I've been locally testing using variants of the configuration below:

Configuration

shard0_count=1
shard0_start=0
shard0_end=255
shard0_0_endpoint="0.0.0.0:5002"
shard0_0_raft_endpoint="0.0.0.0:5003"
shard0_0_readonly_endpoint="0.0.0.0:5004"
shard0_audit_log="shard0_audit_log"
shard1_count=1
shard1_start=128
shard1_end=255
shard1_0_endpoint="0.0.0.0:5005"
shard1_0_raft_endpoint="0.0.0.0:5006"
shard1_0_readonly_endpoint="0.0.0.0:5007"
shard1_audit_log="shard1_audit_log"
coordinator0_count=1
coordinator0_0_endpoint="0.0.0.0:6001"
coordinator0_0_raft_endpoint="0.0.0.0:6002"
coordinator1_count=1
coordinator1_0_endpoint="0.0.0.0:6003"
coordinator1_0_raft_endpoint="0.0.0.0:6004"
sentinel0_endpoint="0.0.0.0:7003"
sentinel1_endpoint="0.0.0.0:8003"
sentinel2_endpoint="0.0.0.0:9003"
sentinel3_endpoint="0.0.0.0:9004"
coordinator0_loglevel="TRACE"
coordinator1_loglevel="TRACE"
sentinel0_loglevel="TRACE"
sentinel1_loglevel="TRACE"
sentinel2_loglevel="TRACE"
sentinel3_loglevel="TRACE"
shard0_loglevel="WARN"
shard1_loglevel="TRACE"
twophase-gen0_loglevel="TRACE"
loadgen0_loglevel="DEBUG"
stxo_cache_depth=0
window_size=100000
batch_size=3000
watchtower_block_cache_size=100
watchtower_error_cache_size=10000000
audit_interval=900
attestation_threshold=2
target_block_interval=250
election_timeout_upper=1000
election_timeout_lower=500
heartbeat=250
raft_max_batch=100000
snapshot_distance=0
batch_delay=1

seed_privkey="a0f36553548b3a66c003413140d7b59e43464ca11af66f25a6e746be501596b7"
seed_value=10
seed_from=0
seed_to=10000

loadgen_sendtx_input_count=2
loadgen_sendtx_output_count=2
loadgen_invalid_tx_rate=0.000000
loadgen_fixed_tx_rate=0.07
loadgen_tps_target=850
loadgen_tps_step_time=3.0
loadgen_tps_step_percentage=0.05
loadgen_tps_step_start=0.2

shard_count=1
coordinator_count=2
loadgen_count=1
sentinel_count=2

2pc=1
sentinel0_private_key="1ffed20f548336af31ddc58fb8db7ef6195b91d1851021cf9fc3b8b196f58fc0"
sentinel0_public_key="6e4fae7a0c180ff20f54698136e4585bf7be045cd5ad621f5c8881871e29c024"
sentinel1_private_key="a94d4bfed8d8679a97f74964c2e19b07098d291ae4967f8f0417821c18253b0e"
sentinel1_public_key="3f8f5f79191ef88d25ede8453f09cb8e1ae5b7b1d98aece1cb25e9d047eca505"
sentinel2_private_key="1ffed20f548336af31ddc58fb8db7ef6195b91d1851021cf9fc3b8b196f58fc0"
sentinel2_public_key="6e4fae7a0c180ff20f54698136e4585bf7be045cd5ad621f5c8881871e29c024"
sentinel3_private_key="a94d4bfed8d8679a97f74964c2e19b07098d291ae4967f8f0417821c18253b0e"
sentinel3_public_key="3f8f5f79191ef88d25ede8453f09cb8e1ae5b7b1d98aece1cb25e9d047eca505"

[HalosGhost]

@maurermi, this doesn't yet include the atomizer updates; but I'd love a review before I dive into that side.

The types and configuration have been modified to mirror the exact current behavior of the test controller (so any currently deployed test controller should be able to now leverage targeted-load schedules in 2PC with no redeployment). Let me know if you want our TC spun up at any point this next week to give a t-ACK.

[maurermi]

T-ACK. Confirmed that these changes work locally for allowing the 2PC-loadgen to ramp up its workload as specified. There is a noticeable load "spike" at the beginning of the test that we will need to address (as discussed out of band) but as this stands it is a significant improvement to the 2PC load generator that can be used as it stands.

[HalosGhost]

(pulled to draft as I implement equivalent changes for the atomizer)