Support fluctuating task load in matching simulation

[taylanisikdemir]

What changed?

• Define interval based task generation qps config to simulation.

    tasks:
      - numtaskgenerators: 3
        maxtasktogenerate: 8000
        overtime:
        - taskspersecond: 10
          tasksburst: 10
          duration: 15s
        - taskspersecond: 250
          tasksburst: 250
          duration: 30s
        - taskspersecond: 10
          tasksburst: 10

• Use different rate limiter per given time window during simulation
• Make qps tracker's interval configurable. It was hardcoded to 10s which was reacting slow to qps spike generated in simulation
• Emit read/write partition count metrics as 1 when partitionConfig is not set in DB. It's never set unless adaptive scaler decides to upscale which was causing gaps in the metrics.

Why?
To simulate/validate adaptive tasklist partitioner behavior.

How did you test it?
Run simulation

go run tools/matchingsimulationcomparison/*.go --scenarios "fluctuating_adaptive"

Visit matching dashboard at http://localhost:3000.
Check qps per tasklist dashboards and validate partition upscale/downscale occurs

Added r/w partition counts dashboard and checked