Capacity estimation method combined with online business index characteristics
A technology that combines business indicators and lines, applied in the field of IT capacity management, can solve problems such as low degree of automation, low accuracy, and timeliness delay, and achieve the effect of improving the degree of automation, improving accuracy, and improving data
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0089] Example 1, the number of incoming items [67,343,65,87,43] after normalization [0.08,1.0,0.073,0.14,0.0]
[0090] Further, the hardware resource data includes cpu, memory and disk data.
[0091] In this embodiment, the resource data collation result of a certain system is as follows figure 2 shown.
[0092] Among them, step 2 specifically includes:
[0093] Step 2.1: The resource monitoring agent tool collects the business indicator data into the data warehouse, and then uses the etl tool to extract the business indicator data (incoming shipments, loans and repayments) in the data warehouse and sort them by time;
[0094] Step 2.2: Since the data such as the incoming quantity, the number of loans, and the number of repayments are not of the same magnitude, the sorted business index data is processed into data ranging from 0 to 1 using the deviation standardization method, which can also make subsequent model training The process avoids weight inclination...
Embodiment 2
[0102] Embodiment 2, business index [0.08, 1.0, 0.073, 0.14, 0.0], cpu data [0.1, 0.2, 0.3, 0.4, 0.6]
[0103] cor=-0.396943, indicating that the business growth trend is negatively correlated with this system
[0104] Among them, step 4 specifically includes:
[0105] Step 4.1: Smooth the cleaned business indicator data to eliminate data noise;
[0106] Step 4.2: Use the prophet time series model to fit and model the smoothed business indicator data, and define the lower limit of the historical data of the business indicator data in the past 30 days as ylower;
[0107] Step 4.3: The integrated historical data is listed as X=, and the forecast standard data is Y= , and take X, Y as the historical training data, and correspond one-to-one in order.
Embodiment 3
[0108] Embodiment three: X=[0.1,0.1,0.1], [0.2,0.2,0.2], [0.3,0.3,0.3],]
[0109] Y = [15.3, 36.9, 55.5]
[0110] Further, the modeling mainly includes the following steps:
[0111] Step 4.2.1: Build python3 and fbprophet environment;
[0112] Step 4.2.2: Introduce the fbprophet package, and call the Prophet method in the fbprophet package, select the kernel function as "linear", set the holiday date of the coming year for holidays, and set the prediction width to 0.5;
[0113] Step 4.2.3: Call the fit method, and input the smoothed business indicator data into the function as a parameter in the standard format;
[0114] Step 4.2.4: Call the make_future_dataframe method, select the forecast period as 30, the date unit as 'd', and define the lower limit of the historical data of the business indicator data in the past 30 days as ylower.
[0115] Further, the pseudocode used in the step 4 time series model is as follows:
[0116] Algorithm time series
[...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More