1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
|
// Copyright Verizon Media. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
package com.yahoo.vespa.hosted.provision.autoscale;
import com.yahoo.vespa.hosted.provision.applications.Application;
import java.time.Duration;
/**
* A resource target to hit for the allocation optimizer.
* The target is measured in cpu, memory and disk per node in the allocation given by current.
*
* @author bratseth
*/
public class ResourceTarget {
private final boolean adjustForRedundancy;
/** The target real resources per node, assuming the node assignment where this was decided */
private final double cpu, memory, disk;
private ResourceTarget(double cpu, double memory, double disk, boolean adjustForRedundancy) {
this.cpu = cpu;
this.memory = memory;
this.disk = disk;
this.adjustForRedundancy = adjustForRedundancy;
}
/** Are the target resources given by this including redundancy or not */
public boolean adjustForRedundancy() { return adjustForRedundancy; }
/** Returns the target cpu per node, in terms of the current allocation */
public double nodeCpu() { return cpu; }
/** Returns the target memory per node, in terms of the current allocation */
public double nodeMemory() { return memory; }
/** Returns the target disk per node, in terms of the current allocation */
public double nodeDisk() { return disk; }
@Override
public String toString() {
return "target " +
(adjustForRedundancy ? "(with redundancy adjustment) " : "") +
"[vcpu " + cpu + ", memoryGb " + memory + ", diskGb " + disk + "]";
}
private static double nodeUsage(Resource resource, double load, AllocatableClusterResources current) {
return load * resource.valueFrom(current.realResources().nodeResources());
}
/** Create a target of achieving ideal load given a current load */
public static ResourceTarget idealLoad(ClusterTimeseries clusterTimeseries,
ClusterNodesTimeseries clusterNodesTimeseries,
AllocatableClusterResources current,
Application application) {
return new ResourceTarget(nodeUsage(Resource.cpu, clusterNodesTimeseries.averageLoad(Resource.cpu), current)
/ idealCpuLoad(clusterTimeseries, clusterNodesTimeseries, application),
nodeUsage(Resource.memory, clusterNodesTimeseries.averageLoad(Resource.memory), current)
/ Resource.memory.idealAverageLoad(),
nodeUsage(Resource.disk, clusterNodesTimeseries.averageLoad(Resource.disk), current)
/ Resource.disk.idealAverageLoad(),
true);
}
/** Crete a target of preserving a current allocation */
public static ResourceTarget preserve(AllocatableClusterResources current) {
return new ResourceTarget(current.realResources().nodeResources().vcpu(),
current.realResources().nodeResources().memoryGb(),
current.realResources().nodeResources().diskGb(),
false);
}
/** Ideal cpu load must take the application traffic fraction into account */
private static double idealCpuLoad(ClusterTimeseries clusterTimeseries,
ClusterNodesTimeseries clusterNodesTimeseries,
Application application) {
// What's needed to have headroom for growth during scale-up as a fraction of current resources?
double maxGrowthRate = clusterTimeseries.maxQueryGrowthRate(); // in fraction per minute of the current traffic
Duration scalingDuration = clusterNodesTimeseries.cluster().scalingDuration(clusterNodesTimeseries.clusterNodes().clusterSpec());
double growthRateHeadroom = 1 + maxGrowthRate * scalingDuration.toMinutes();
// Cap headroom at 10% above the historical observed peak
growthRateHeadroom = Math.min(growthRateHeadroom, 1 / clusterTimeseries.currentQueryFractionOfMax() + 0.1);
// How much headroom is needed to handle sudden arrival of additional traffic due to another zone going down?
double trafficShiftHeadroom;
if (application.status().maxReadShare() == 0) // No traffic fraction data
trafficShiftHeadroom = 2.0; // assume we currently get half of the global share of traffic
else
trafficShiftHeadroom = application.status().maxReadShare() / application.status().currentReadShare();
if (trafficShiftHeadroom > 2.0) // The expectation that we have almost no load with almost no queries is incorrect due
trafficShiftHeadroom = 2.0; // to write traffic; once that is separated we can increase this threshold
return 1 / growthRateHeadroom * 1 / trafficShiftHeadroom * Resource.cpu.idealAverageLoad();
}
}
|
