Update docs: consistent-hashing sharding algorithm

cmd/argocd-application-controller/commands/argocd_application_controller.go

@@ -220,7 +220,7 @@ func NewCommand() *cobra.Command {
 	command.Flags().StringSliceVar(&otlpAttrs, "otlp-attrs", env.StringsFromEnv("ARGOCD_APPLICATION_CONTROLLER_OTLP_ATTRS", []string{}, ","), "List of OpenTelemetry collector extra attrs when send traces, each attribute is separated by a colon(e.g. key:value)")
 	command.Flags().StringSliceVar(&applicationNamespaces, "application-namespaces", env.StringsFromEnv("ARGOCD_APPLICATION_NAMESPACES", []string{}, ","), "List of additional namespaces that applications are allowed to be reconciled from")
 	command.Flags().BoolVar(&persistResourceHealth, "persist-resource-health", env.ParseBoolFromEnv("ARGOCD_APPLICATION_CONTROLLER_PERSIST_RESOURCE_HEALTH", true), "Enables storing the managed resources health in the Application CRD")
-	command.Flags().StringVar(&shardingAlgorithm, "sharding-method", env.StringFromEnv(common.EnvControllerShardingAlgorithm, common.DefaultShardingAlgorithm), "Enables choice of sharding method. Supported sharding methods are : [legacy, round-robin] ")
+	command.Flags().StringVar(&shardingAlgorithm, "sharding-method", env.StringFromEnv(common.EnvControllerShardingAlgorithm, common.DefaultShardingAlgorithm), "Enables choice of sharding method. Supported sharding methods are : [legacy, round-robin, consistent-hashing] ")
 	// global queue rate limit config
 	command.Flags().Int64Var(&workqueueRateLimit.BucketSize, "wq-bucket-size", env.ParseInt64FromEnv("WORKQUEUE_BUCKET_SIZE", 500, 1, math.MaxInt64), "Set Workqueue Rate Limiter Bucket Size, default 500")
 	command.Flags().Float64Var(&workqueueRateLimit.BucketQPS, "wq-bucket-qps", env.ParseFloat64FromEnv("WORKQUEUE_BUCKET_QPS", math.MaxFloat64, 1, math.MaxFloat64), "Set Workqueue Rate Limiter Bucket QPS, default set to MaxFloat64 which disables the bucket limiter")

cmd/argocd/commands/admin/cluster.go

@@ -219,7 +219,7 @@ func NewClusterShardsCommand(clientOpts *argocdclient.ClientOptions) *cobra.Comm
 	clientConfig = cli.AddKubectlFlagsToCmd(&command)
 	command.Flags().IntVar(&shard, "shard", -1, "Cluster shard filter")
 	command.Flags().IntVar(&replicas, "replicas", 0, "Application controller replicas count. Inferred from number of running controller pods if not specified")
-	command.Flags().StringVar(&shardingAlgorithm, "sharding-method", common.DefaultShardingAlgorithm, "Sharding method. Defaults: legacy. Supported sharding methods are : [legacy, round-robin] ")
+	command.Flags().StringVar(&shardingAlgorithm, "sharding-method", common.DefaultShardingAlgorithm, "Sharding method. Defaults: legacy. Supported sharding methods are : [legacy, round-robin, consistent-hashing] ")
 	command.Flags().BoolVar(&portForwardRedis, "port-forward-redis", true, "Automatically port-forward ha proxy redis from current namespace?")
 
 	cacheSrc = appstatecache.AddCacheFlagsToCmd(&command)
@@ -514,7 +514,7 @@ argocd admin cluster stats target-cluster`,
 	clientConfig = cli.AddKubectlFlagsToCmd(&command)
 	command.Flags().IntVar(&shard, "shard", -1, "Cluster shard filter")
 	command.Flags().IntVar(&replicas, "replicas", 0, "Application controller replicas count. Inferred from number of running controller pods if not specified")
-	command.Flags().StringVar(&shardingAlgorithm, "sharding-method", common.DefaultShardingAlgorithm, "Sharding method. Defaults: legacy. Supported sharding methods are : [legacy, round-robin] ")
+	command.Flags().StringVar(&shardingAlgorithm, "sharding-method", common.DefaultShardingAlgorithm, "Sharding method. Defaults: legacy. Supported sharding methods are : [legacy, round-robin, consistent-hashing] ")
 	command.Flags().BoolVar(&portForwardRedis, "port-forward-redis", true, "Automatically port-forward ha proxy redis from current namespace?")
 	cacheSrc = appstatecache.AddCacheFlagsToCmd(&command)
 

docs/operator-manual/high_availability.md

@@ -82,10 +82,16 @@ spec:
 ```
 * In order to manually set the cluster's shard number, specify the optional `shard` property when creating a cluster. If not specified, it will be calculated on the fly by the application controller.
 
-* The shard distribution algorithm of the `argocd-application-controller` can be set by using the `--sharding-method` parameter. Supported sharding methods are : [legacy (default), round-robin]. `legacy` mode uses an `uid` based distribution (non-uniform). `round-robin` uses an equal distribution across all shards. The `--sharding-method` parameter can also be overridden by setting the key `controller.sharding.algorithm` in the `argocd-cmd-params-cm` `configMap` (preferably) or by setting the `ARGOCD_CONTROLLER_SHARDING_ALGORITHM` environment variable and by specifiying the same possible values.
+* The shard distribution algorithm of the `argocd-application-controller` can be set by using the `--sharding-method` parameter. Supported sharding methods are : [legacy (default), round-robin, consistent-hashing]:
+- `legacy` mode uses an `uid` based distribution (non-uniform).
+- `round-robin` uses an equal distribution across all shards.
+- `consistent-hashing` uses the consistent hashing with bounded loads algorithm which tends to equal distribution and also reduces cluster or application reshuffling in case of additions or removals of shards or clusters. 
 
-!!! warning "Alpha Feature"
-    The `round-robin` shard distribution algorithm  is an experimental feature. Reshuffling is known to occur in certain scenarios with cluster removal. If the cluster at rank-0 is removed, reshuffling all clusters across shards will occur and may temporarily have negative performance impacts.
+The `--sharding-method` parameter can also be overridden by setting the key `controller.sharding.algorithm` in the `argocd-cmd-params-cm` `configMap` (preferably) or by setting the `ARGOCD_CONTROLLER_SHARDING_ALGORITHM` environment variable and by specifiying the same possible values.
+
+!!! warning "Alpha Features"
+    The `round-robin` shard distribution algorithm is an experimental feature. Reshuffling is known to occur in certain scenarios with cluster removal. If the cluster at rank-0 is removed, reshuffling all clusters across shards will occur and may temporarily have negative performance impacts.
+    The `consistent-hashing` shard distribution algorithm is an experimental feature. Extensive benchmark have been documented on the [CNOE blog](https://cnoe.io/blog/argo-cd-application-scalability) with encouraging results. Community feedback is highly appreciated before moving this feature to a production ready state.
 
 * A cluster can be manually assigned and forced to a `shard` by patching the `shard` field in the cluster secret to contain the shard number, e.g.
 ```yaml