public static Recommendation makeRecommendation()

in src/main/java/com/google/cloud/run/kafkascaler/LagScaling.java [53:78]

23 lines of code
3 McCabe index (conditional complexity)


  public static Recommendation makeRecommendation(
      MetricTarget lagTarget, int currentInstanceCount, long currentLag) {
    if (currentLag <= lagTarget.activationThreshold()) {
      logger.atInfo().log(
          "[LAG] Scaling inactive, current lag: %d, activation lag threshold: %d",
          currentLag, lagTarget.activationThreshold());
      return new Recommendation(false, 0);
    }

    int recommendedInstanceCount = currentInstanceCount;
    double scalingFactor = currentLag / (double) lagTarget.averageValue();
    double upperTolerance = (1 + lagTarget.tolerance()) * lagTarget.averageValue();
    double lowerTolerance = (1 - lagTarget.tolerance()) * lagTarget.averageValue();
    logger.atInfo().log(
        "[LAG] Current Lag: %d, upper tolerance: %.2f, lower tolerance: %.2f\n",
        currentLag, upperTolerance, lowerTolerance);

    if (currentLag <= upperTolerance && currentLag >= lowerTolerance) {
      logger.atInfo().log("[LAG] Within tolerance, no change");
    } else {
      recommendedInstanceCount = (int) Math.ceil(max(currentInstanceCount, 1) * scalingFactor);
      logger.atInfo().log("[LAG] Recommended instance count: %d", recommendedInstanceCount);
    }

    return new Recommendation(true, recommendedInstanceCount);
  }