2
apiVersion: apiextensions.k8s.io/v1
3
kind: CustomResourceDefinition
6
controller-gen.kubebuilder.io/version: v0.14.0
7
name: meshloadbalancingstrategies.kuma.io
13
kind: MeshLoadBalancingStrategy
14
listKind: MeshLoadBalancingStrategyList
15
plural: meshloadbalancingstrategies
16
singular: meshloadbalancingstrategy
19
- additionalPrinterColumns:
20
- jsonPath: .spec.targetRef.kind
23
- jsonPath: .spec.targetRef.name
32
APIVersion defines the versioned schema of this representation of an object.
33
Servers should convert recognized schemas to the latest internal value, and
34
may reject unrecognized values.
35
More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
39
Kind is a string value representing the REST resource this object represents.
40
Servers may infer this from the endpoint the client submits requests to.
43
More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
48
description: Spec is the specification of the Kuma MeshLoadBalancingStrategy
53
TargetRef is a reference to the resource the policy takes an effect on.
54
The resource could be either a real store object or virtual resource
58
description: Kind of the referenced resource
68
description: Mesh is reserved for future use to identify cross
73
Name of the referenced resource. Can only be used with kinds: `MeshService`,
74
`MeshServiceSubset` and `MeshGatewayRoute`
78
ProxyTypes specifies the data plane types that are subject to the policy. When not specified,
79
all data plane types are targeted by the policy.
91
Tags used to select a subset of proxies by tags. Can only be used with kinds
92
`MeshSubset` and `MeshServiceSubset`
96
description: To list makes a match between the consumed services and
97
corresponding configurations
102
Default is a configuration specific to the group of destinations referenced in
106
description: LoadBalancer allows to specify load balancing
111
LeastRequest selects N random available hosts as specified in 'choiceCount' (2 by default)
112
and picks the host which has the fewest active requests
119
ActiveRequestBias refers to dynamic weights applied when hosts have varying load
120
balancing weights. A higher value here aggressively reduces the weight of endpoints
121
that are currently handling active requests. In essence, the higher the ActiveRequestBias
122
value, the more forcefully it reduces the load balancing weight of endpoints that are
123
actively serving requests.
124
x-kubernetes-int-or-string: true
127
ChoiceCount is the number of random healthy hosts from which the host with
128
the fewest active requests will be chosen. Defaults to 2 so that Envoy performs
129
two-choice selection if the field is not set.
136
Maglev implements consistent hashing to upstream hosts. Maglev can be used as
137
a drop in replacement for the ring hash load balancer any place in which
138
consistent hashing is desired.
142
HashPolicies specify a list of request/connection properties that are used to calculate a hash.
143
These hash policies are executed in the specified order. If a hash policy has the “terminal” attribute
144
set to true, and there is already a hash generated, the hash is returned immediately,
145
ignoring the rest of the hash policy list.
151
description: Hash on source IP address.
157
description: The name of the cookie that
158
will be used to obtain the hash key.
162
description: The name of the path for
166
description: If specified, a cookie with
167
the TTL will be generated if the cookie
177
The name of the Object in the per-request filterState, which is
178
an Envoy::Hashable object. If there is no data associated with the key,
179
or the stored object is not Envoy::Hashable, no hash will be produced.
188
description: The name of the request header
189
that will be used to obtain the hash
200
The name of the URL query parameter that will be used to obtain the hash key.
201
If the parameter is not present, no hash will be produced. Query parameter names
210
Terminal is a flag that short-circuits the hash computing. This field provides
211
a ‘fallback’ style of configuration: “if a terminal policy doesn’t work, fallback
212
to rest of the policy list”, it saves time when the terminal policy works.
213
If true, and there is already a hash computed, ignore rest of the list of hash polices.
229
The table size for Maglev hashing. Maglev aims for “minimal disruption”
230
rather than an absolute guarantee. Minimal disruption means that when
231
the set of upstream hosts change, a connection will likely be sent
232
to the same upstream as it was before. Increasing the table size reduces
233
the amount of disruption. The table size must be prime number limited to 5000011.
234
If it is not specified, the default is 65537.
242
Random selects a random available host. The random load balancer generally
243
performs better than round-robin if no health checking policy is configured.
244
Random selection avoids bias towards the host in the set that comes after a failed host.
248
RingHash implements consistent hashing to upstream hosts. Each host is mapped
249
onto a circle (the “ring”) by hashing its address; each request is then routed
250
to a host by hashing some property of the request, and finding the nearest
251
corresponding host clockwise around the ring.
255
HashFunction is a function used to hash hosts onto the ketama ring.
256
The value defaults to XX_HASH. Available values – XX_HASH, MURMUR_HASH_2.
263
HashPolicies specify a list of request/connection properties that are used to calculate a hash.
264
These hash policies are executed in the specified order. If a hash policy has the “terminal” attribute
265
set to true, and there is already a hash generated, the hash is returned immediately,
266
ignoring the rest of the hash policy list.
272
description: Hash on source IP address.
278
description: The name of the cookie that
279
will be used to obtain the hash key.
283
description: The name of the path for
287
description: If specified, a cookie with
288
the TTL will be generated if the cookie
298
The name of the Object in the per-request filterState, which is
299
an Envoy::Hashable object. If there is no data associated with the key,
300
or the stored object is not Envoy::Hashable, no hash will be produced.
309
description: The name of the request header
310
that will be used to obtain the hash
321
The name of the URL query parameter that will be used to obtain the hash key.
322
If the parameter is not present, no hash will be produced. Query parameter names
331
Terminal is a flag that short-circuits the hash computing. This field provides
332
a ‘fallback’ style of configuration: “if a terminal policy doesn’t work, fallback
333
to rest of the policy list”, it saves time when the terminal policy works.
334
If true, and there is already a hash computed, ignore rest of the list of hash polices.
350
Maximum hash ring size. Defaults to 8M entries, and limited to 8M entries,
351
but can be lowered to further constrain resource use.
358
Minimum hash ring size. The larger the ring is (that is,
359
the more hashes there are for each provided host) the better the request distribution
360
will reflect the desired weights. Defaults to 1024 entries, and limited to 8M entries.
368
RoundRobin is a load balancing algorithm that distributes requests
369
across available upstream hosts in round-robin order.
383
description: LocalityAwareness contains configuration for
384
locality aware load balancing.
388
CrossZone defines locality aware load balancing priorities when dataplane proxies inside local zone
392
description: Failover defines list of load balancing
393
rules in order of priority
397
description: From defines the list of zones
398
to which the rule applies
408
description: To defines to which zones the
409
traffic should be load balanced
412
description: Type defines how target zones
413
will be picked from available zones
433
FailoverThreshold defines the percentage of live destination dataplane proxies below which load balancing to the
434
next priority starts.
435
Example: If you configure failoverThreshold to 70, and you have deployed 10 destination dataplane proxies.
436
Load balancing to next priority will start when number of live destination dataplane proxies drops below 7.
443
x-kubernetes-int-or-string: true
450
Disabled allows to disable locality-aware load balancing.
451
When disabled requests are distributed across all endpoints regardless of locality.
454
description: LocalZone defines locality aware load balancing
455
priorities between dataplane proxies inside a zone
458
description: AffinityTags list of tags for local
463
description: Key defines tag for which affinity
468
Weight of the tag used for load balancing. The bigger the weight the bigger the priority.
469
Percentage of local traffic load balanced to tag is computed by dividing weight by sum of weights from all tags.
470
For example with two affinity tags first with weight 80 and second with weight 20,
471
then 80% of traffic will be redirected to the first tag, and 20% of traffic will be redirected to second one.
472
Setting weights is not mandatory. When weights are not set control plane will compute default weight based on list order.
473
Default: If you do not specify weight we will adjust them so that 90% traffic goes to first tag, 9% to next, and 1% to third and so on.
485
TargetRef is a reference to the resource that represents a group of
489
description: Kind of the referenced resource
499
description: Mesh is reserved for future use to identify
500
cross mesh resources.
504
Name of the referenced resource. Can only be used with kinds: `MeshService`,
505
`MeshServiceSubset` and `MeshGatewayRoute`
509
ProxyTypes specifies the data plane types that are subject to the policy. When not specified,
510
all data plane types are targeted by the policy.
519
additionalProperties:
522
Tags used to select a subset of proxies by tags. Can only be used with kinds
523
`MeshSubset` and `MeshServiceSubset`