add HMS PodDisruption Budget (#280)

* add HMS PodDisruption Budget make Rolling Update Strategy settings for HMS RO and RW configurable * Update CHANGELOG.md --------- Co-authored-by: Georgi Ivanov <[email protected]> Co-authored-by: Patrick Duin <[email protected]>
ExpediaGroup · Oct 15, 2024 · 062f300 · 062f300
1 parent 2052e52
commit 062f300
Show file tree

Hide file tree

Showing 6 changed files with 131 additions and 2 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -3,6 +3,11 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/) and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
+## [7.5.0] - 2024-10-15
+### Added
+- Added variables `hms_rw_k8s_pdb_settings` and `hms_ro_k8s_pdb_settings` to specify HMS ro and rw PodDisruptionBudget. Uses policy/v1 version which is evailable since kubernetes 1.25+
+- Added variables `hms_rw_k8s_rolling_update_strategy` and `hms_ro_k8s_rolling_update_strategy` to specify Deployment rolling update strategy settings for HMS ro and rw pods.
+
 ## [7.4.0] - 2024-09-25
 ### Added
 - Added variables `hms_rw_tolerations` and `hms_ro_tolerations` to specify tolerations for the HMS ro and rw pods

diff --git a/VARIABLES.md b/VARIABLES.md
@@ -70,6 +70,8 @@
 | hms\_ro\_heapsize                                         | Heapsize for the read only Hive Metastore.<br>Valid values: https://docs.aws.amazon.com/AmazonECS/latest/developerguide/task-cpu-memory-error.html                                                                   | `string`                 | `"2048"`                                                                                                                                                                                                                                                                                       |    no    |
 | hms\_ro\_k8s\_replica\_count                              | Initial Number of read only Hive Metastore k8s pod replicas to create.                                                                                                                                               | `number`                 | `"2048"`                                                                                                                                                                                                                                                                                       |    no    |
 | hms\_ro\_k8s\_max\_replica\_count                         | Max Number of read only Hive Metastore k8s pod replicas to create.                                                                                                                                                   | `number`                 | `"2048"`                                                                                                                                                                                                                                                                                       |    no    |
+| hms\_rw\_k8s\_pdb\_settings                        | Add PodDisruptionBudget to the HMS rw pods.                                                                                                                                                              | `object`                 | `max_unavailable = 1`                                                                                                                                                                                                                                                                                            |    no    |
+| hms\_rw\_k8s\_rolling\_update\_strategy                        | Configure HMS RW deployment rolling strategy.                                                                                                                                                              | `object`                 | `max_unavailable = 1`                                                                                                                                                                                                                                                                                            |    no    |
 | hms\_ro\_target\_cpu\_percentage                          | Read only Hive Metastore autoscaling threshold for CPU target usage.                                                                                                                                                 | `number`                 | `"2048"`                                                                                                                                                                                                                                                                                       |    no    |
 | hms\_ro\_request\_partition\_limit                        | Read only Hive Metastore limits of request partitions.                                                                                                                                                               | `string`                 | n/a                                                                                                                                                                                                                                                                                            |    no    |
 | hms\_ro\_node\_affinity                        | Add node affinities to the Hive metastore pods.                                                                                                                                                              | `list(object)`                 | n/a                                                                                                                                                                                                                                                                                            |    no    |
@@ -79,6 +81,8 @@
 | hms\_rw\_ecs\_task\_count                                 | Desired ECS task count of the read/write Hive Metastore service.                                                                                                                                                     | `string`                 | `"3"`                                                                                                                                                                                                                                                                                          |    no    |
 | hms\_rw\_heapsize                                         | Heapsize for the read/write Hive Metastore.<br>Valid values: https://docs.aws.amazon.com/AmazonECS/latest/developerguide/task-cpu-memory-error.html                                                                  | `string`                 | `"2048"`                                                                                                                                                                                                                                                                                       |    no    |
 | hms\_rw\_k8s\_replica\_count                              | Initial Number of read/write Hive Metastore k8s pod replicas to create.                                                                                                                                              | `number`                 | `"2048"`                                                                                                                                                                                                                                                                                       |    no    |
+| hms\_rw\_k8s\_pdb\_settings                        | Add PodDisruptionBudget to the HMS rw pods.                                                                                                                                                              | `object`                 | `max_unavailable = 1`                                                                                                                                                                                                                                                                                            |    no    |
+| hms\_rw\_k8s\_rolling\_update\_strategy                        | Configure HMS RW deployment rolling strategy.                                                                                                                                                              | `object`                 | `max_unavailable = 1`                                                                                                                                                                                                                                                                                            |    no    |
 | hms\_rw\_request\_partition\_limit                        | Read Write Hive Metastore limits of request partitions.                                                                                                                                                              | `string`                 | n/a                                                                                                                                                                                                                                                                                            |    no    |
 | hms\_rw\_node\_affinity                        | Add node affinities to the Hive metastore pods.                                                                                                                                                              | `list(object)`                 | n/a                                                                                                                                                                                                                                                                                            |    no    |
 | hms\_rw\_tolerations                        | Add tolerations to the Hive metastore pods.                                                                                                                                                              | `list(object)`                 | n/a                                                                                                                                                                                                                                                                                            |    no    |
@@ -356,4 +360,4 @@ apiary_managed_schemas = [
    producer_roles = "arn:aws:iam::000000000:role/role-1,arn:aws:iam::000000000:role/role-2"
   }
 ]
-```
+```
diff --git a/k8s-readonly.tf b/k8s-readonly.tf
@@ -16,6 +16,13 @@ resource "kubernetes_deployment_v1" "apiary_hms_readonly" {
   }
 
   spec {
+    strategy {
+      type = "RollingUpdate"
+      rolling_update {
+        max_surge       = var.hms_ro_k8s_rolling_update_strategy.max_surge
+        max_unavailable = var.hms_ro_k8s_rolling_update_strategy.max_unavailable
+      }
+    }
     replicas = var.hms_ro_k8s_replica_count
     selector {
       match_labels = {
@@ -348,3 +355,24 @@ data "aws_lb" "k8s_hms_ro_lb" {
   count = var.hms_instance_type == "k8s" && var.enable_vpc_endpoint_services ? 1 : 0
   name  = split("-", split(".", kubernetes_service.hms_readonly[0].status.0.load_balancer.0.ingress.0.hostname).0).0
 }
+
+resource "kubernetes_pod_disruption_budget_v1" "hms_readonly" {
+  count = var.hms_instance_type == "k8s" && var.hms_ro_k8s_pdb_settings.enabled ? 1 : 0
+
+  metadata {
+    name = "${local.hms_alias}-readonly"
+    namespace = var.metastore_namespace
+  }
+
+  spec {
+    selector {
+      match_labels = {
+        name = "${local.hms_alias}-readonly"
+      }
+    }
+
+    # set max_unavailable to 1 by default if PDB is created
+    max_unavailable = var.hms_ro_k8s_pdb_settings.max_unavailable != null ? var.hms_ro_k8s_pdb_settings.max_unavailable : "1"
+    min_available   = var.hms_ro_k8s_pdb_settings.min_available != null ? var.hms_ro_k8s_pdb_settings.min_available : null
+  }
+}
diff --git a/k8s-readwrite.tf b/k8s-readwrite.tf
@@ -16,6 +16,13 @@ resource "kubernetes_deployment_v1" "apiary_hms_readwrite" {
   }
 
   spec {
+    strategy {
+      type = "RollingUpdate"
+      rolling_update {
+        max_surge       = var.hms_rw_k8s_rolling_update_strategy.max_surge
+        max_unavailable = var.hms_rw_k8s_rolling_update_strategy.max_unavailable
+      }
+    }    
     replicas = var.hms_rw_k8s_replica_count
     selector {
       match_labels = {
@@ -366,3 +373,24 @@ data "aws_lb" "k8s_hms_rw_lb" {
   count = var.hms_instance_type == "k8s" && var.enable_vpc_endpoint_services ? 1 : 0
   name  = split("-", split(".", kubernetes_service.hms_readwrite[0].status.0.load_balancer.0.ingress.0.hostname).0).0
 }
+
+resource "kubernetes_pod_disruption_budget_v1" "hms_readwrite" {
+  count = var.hms_instance_type == "k8s" && var.hms_rw_k8s_pdb_settings.enabled ? 1 : 0
+
+  metadata {
+    name = "${local.hms_alias}-readwrite"
+    namespace = var.metastore_namespace
+  }
+
+  spec {
+    selector {
+      match_labels = {
+        name = "${local.hms_alias}-readwrite"
+      }
+    }
+
+    # set max_unavailable to 1 by default if PDB is created
+    max_unavailable = var.hms_rw_k8s_pdb_settings.max_unavailable != null ? var.hms_rw_k8s_pdb_settings.max_unavailable : "1"
+    min_available   = var.hms_rw_k8s_pdb_settings.min_available != null ? var.hms_rw_k8s_pdb_settings.min_available : null
+  }
+}
diff --git a/variables.tf b/variables.tf
@@ -394,6 +394,58 @@ variable "hms_ro_k8s_max_replica_count" {
   default     = 10
 }
 
+variable "hms_ro_k8s_rolling_update_strategy" {
+  description = "Rolling update strategy settings for HMS RO including max_surge and max_unavailable"
+  type = object({
+    max_surge       = string
+    max_unavailable = string
+  })
+  default = {
+    max_surge       = "25%"
+    max_unavailable = "25%"
+  }
+}
+
+variable "hms_rw_k8s_rolling_update_strategy" {
+  description = "Rolling update strategy settings for HMS RW including max_surge and max_unavailable"
+  type = object({
+    max_surge       = string
+    max_unavailable = string
+  })
+  default = {
+    max_surge       = "25%"
+    max_unavailable = "25%"
+  }
+}
+
+variable "hms_ro_k8s_pdb_settings" {
+  description = "PDB settings for HMS RO including enable flag, maxUnavailable, and minAvailable."
+  type = object({
+    enabled         = bool
+    max_unavailable = string
+    min_available   = string
+  })
+  default = {
+    enabled         = false
+    max_unavailable = null
+    min_available   = null
+  }
+}
+
+variable "hms_rw_k8s_pdb_settings" {
+  description = "PDB settings for HMS RW including enable flag, maxUnavailable, and minAvailable."
+  type = object({
+    enabled         = bool
+    max_unavailable = string
+    min_available   = string
+  })
+  default = {
+    enabled         = false
+    max_unavailable = null
+    min_available   = null
+  }
+}
+
 variable "hms_rw_node_affinity" {
   description = <<EOF
 Adds a list of node affinities for the HMS readwrite pods. For example if you

diff --git a/version.tf b/version.tf
@@ -15,9 +15,21 @@ terraform {
       source  = "hashicorp/aws"
       version = "~> 4.0"
     }
+    external = {
+      source  = "hashicorp/external"
+      version = "~> 1.1"
+    }
+    random = {
+      source  = "hashicorp/random"
+      version = "~> 3.6"
+    }
     datadog = {
-      source = "DataDog/datadog"
+      source  = "datadog/datadog"
       version = "3.25.0"
     }
+    template = {
+      source = "hashicorp/template"
+      version = "~> 2.2"
+    }
   }
 }