feat: update images and screenshots

Author: jvdtoorn

README.md

@@ -26,6 +26,9 @@ make up       # Start services
 - **Dashboard:** http://localhost:3000 (Grafana)
 - **API:** http://localhost:3001 (OpenAPI docs)
 
+![OpenDT Grafana Dashboard](site/src/components/HomepageFeatures/grafana-dashboard.png)
+*Real-time dashboard comparing actual vs. simulated power consumption*
+
 ---
 
 ## Repository Structure

config/default.yaml

@@ -11,6 +11,7 @@ services:
 
   simulator:
     simulation_frequency_minutes: 5
+    background_load_nodes: 0
 
 kafka:
   topics:

config/experiments/experiment_1.yaml

@@ -11,6 +11,7 @@ services:
 
   simulator:
     simulation_frequency_minutes: 60
+    background_load_nodes: 0
 
 kafka:
   topics:

config/experiments/experiment_2.yaml

@@ -11,6 +11,7 @@ services:
 
   simulator:
     simulation_frequency_minutes: 60
+    background_load_nodes: 0
 
   calibrator:
     calibrated_property: "cpuPowerModel.calibrationFactor"

libs/common/odt_common/config.py

@@ -66,6 +66,11 @@ class SimulatorConfig(BaseModel):
     simulation_frequency_minutes: int = Field(
         default=15, description="Simulation frequency in minutes (simulated time)", gt=0
     )
+    background_load_nodes: int = Field(
+        ...,
+        description="Number of nodes reserved for background load (not available for simulation)",
+        ge=0,
+    )
 
 
 class CalibratorConfig(BaseModel):

reproducibility-capsule/output/experiment_1_SURF.pdf

@@ -79,11 +79,18 @@ Settings under `services.simulator` in the config file:
 | Setting | Type | Default | Description |
 |---------|------|---------|-------------|
 | simulation_frequency_minutes | int | 5 | Window size for task aggregation |
+| background_load_nodes | int | (required) | Number of nodes reserved for background load |
 
 ### simulation_frequency_minutes
 
 Determines the time window size for batching tasks. Larger windows mean fewer OpenDC invocations but less granular results.
 
+### background_load_nodes
+
+Simulates a certain number of nodes being used for background load. These nodes are subtracted from the first host type in the first cluster of the topology and are not available for assigning workload in the simulation.
+
+This is useful when the real datacenter has nodes running workloads that are not part of the simulated task trace (e.g., system services, other tenants). Set to `0` to use the full topology.
+
 ## Logs
 
 ```

services/simulator/README.md

@@ -42,6 +42,7 @@ def __init__(
         speed_factor: float,
         run_output_dir: str,
         run_id: str,
+        background_load_nodes: int = 0,
         consumer_group: str = "simulators",
     ):
         """Initialize the simulation service.
@@ -55,6 +56,7 @@ def __init__(
             speed_factor: Configured simulation speed multiplier
             run_output_dir: Base directory for run outputs
             run_id: Unique run ID for this session
+            background_load_nodes: Number of nodes reserved for background load
             consumer_group: Kafka consumer group ID
         """
         self.kafka_bootstrap_servers = kafka_bootstrap_servers
@@ -65,6 +67,7 @@ def __init__(
         self.simulation_frequency = timedelta(minutes=simulation_frequency_minutes)
         self.speed_factor = speed_factor
         self.run_id = run_id
+        self.background_load_nodes = background_load_nodes
 
         # Setup output directories - simulator writes to run_dir/simulator/
         self.output_base_dir = Path(run_output_dir) / run_id / "simulator"
@@ -121,6 +124,50 @@ def __init__(
         logger.info(
             f"Simulation frequency: {simulation_frequency_minutes} minutes (simulated time)"
         )
+        logger.info(f"Background load nodes: {background_load_nodes}")
+
+    def _reduce_topology_for_background_load(self, topology: Topology) -> Topology:
+        """Create a topology with reduced host count to simulate background load.
+
+        Subtracts background_load_nodes from the first host type in the first cluster.
+        This simulates nodes being occupied by background workload and unavailable
+        for the simulation.
+
+        Args:
+            topology: Original topology
+
+        Returns:
+            New topology with reduced host count in first cluster
+        """
+        if self.background_load_nodes == 0:
+            return topology
+
+        # Deep copy to avoid modifying the original
+        reduced = copy.deepcopy(topology)
+
+        if not reduced.clusters or not reduced.clusters[0].hosts:
+            logger.warning("Topology has no clusters or hosts, cannot reduce")
+            return topology
+
+        first_host = reduced.clusters[0].hosts[0]
+        original_count = first_host.count
+
+        if self.background_load_nodes >= original_count:
+            logger.warning(
+                f"background_load_nodes ({self.background_load_nodes}) >= "
+                f"available hosts ({original_count}) in first cluster, "
+                f"setting to {original_count - 1}"
+            )
+            first_host.count = max(1, original_count - self.background_load_nodes)
+        else:
+            first_host.count = original_count - self.background_load_nodes
+
+        logger.info(
+            f"Reduced topology: {original_count} -> {first_host.count} hosts "
+            f"in cluster '{reduced.clusters[0].name}' ({self.background_load_nodes} for background load)"
+        )
+
+        return reduced
 
     def _run_simulation(self) -> None:
         """Run OpenDC simulation with accumulated tasks.
@@ -235,8 +282,8 @@ def _run_simulation(self) -> None:
         # Increment run number
         self.run_number += 1
 
-        # Check if we can reuse cached results
-        topology_to_use = self.simulated_topology
+        # Apply background load reduction to topology
+        topology_to_use = self._reduce_topology_for_background_load(self.simulated_topology)
 
         # Create directories
         run_dir = self.output_base_dir / "opendc" / f"run_{self.run_number}"
@@ -462,6 +509,7 @@ def main():
 
     # Get simulator configuration
     simulation_frequency_minutes = config.services.simulator.simulation_frequency_minutes
+    background_load_nodes = config.services.simulator.background_load_nodes
     speed_factor = config.global_config.speed_factor
     run_output_dir = Path(os.getenv("DATA_DIR", "/app/data"))
 
@@ -470,6 +518,7 @@ def main():
     logger.info(f"Topology topic: {topology_topic}")
     logger.info(f"Simulated topology topic: {sim_topology_topic}")
     logger.info(f"Simulation frequency: {simulation_frequency_minutes} minutes")
+    logger.info(f"Background load nodes: {background_load_nodes}")
     logger.info(f"Speed factor: {speed_factor}x")
     logger.info(f"Data directory: {run_output_dir}")
 
@@ -500,6 +549,7 @@ def main():
                 speed_factor=speed_factor,
                 run_output_dir=str(run_output_dir),
                 run_id=run_id,
+                background_load_nodes=background_load_nodes,
                 consumer_group=consumer_group,
             )
             service.run()