forked from terraform-aws-modules/terraform-aws-eks
-
Notifications
You must be signed in to change notification settings - Fork 0
/
node_groups.tf
562 lines (484 loc) · 37.8 KB
/
node_groups.tf
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
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
locals {
metadata_options = {
http_endpoint = "enabled"
http_tokens = "required"
http_put_response_hop_limit = 2
}
# EKS managed node group
default_update_config = {
max_unavailable_percentage = 33
}
# Self-managed node group
default_instance_refresh = {
strategy = "Rolling"
preferences = {
min_healthy_percentage = 66
}
}
kubernetes_network_config = try(aws_eks_cluster.this[0].kubernetes_network_config[0], {})
}
# This sleep resource is used to provide a timed gap between the cluster creation and the downstream dependencies
# that consume the outputs from here. Any of the values that are used as triggers can be used in dependencies
# to ensure that the downstream resources are created after both the cluster is ready and the sleep time has passed.
# This was primarily added to give addons that need to be configured BEFORE data plane compute resources
# enough time to create and configure themselves before the data plane compute resources are created.
resource "time_sleep" "this" {
count = var.create ? 1 : 0
create_duration = var.dataplane_wait_duration
triggers = {
cluster_name = aws_eks_cluster.this[0].name
cluster_endpoint = aws_eks_cluster.this[0].endpoint
cluster_version = aws_eks_cluster.this[0].version
cluster_service_cidr = var.cluster_ip_family == "ipv6" ? try(local.kubernetes_network_config.service_ipv6_cidr, "") : try(local.kubernetes_network_config.service_ipv4_cidr, "")
cluster_certificate_authority_data = aws_eks_cluster.this[0].certificate_authority[0].data
}
}
################################################################################
# EKS IPV6 CNI Policy
# https://docs.aws.amazon.com/eks/latest/userguide/cni-iam-role.html#cni-iam-role-create-ipv6-policy
################################################################################
data "aws_iam_policy_document" "cni_ipv6_policy" {
count = var.create && var.create_cni_ipv6_iam_policy ? 1 : 0
statement {
sid = "AssignDescribe"
actions = [
"ec2:AssignIpv6Addresses",
"ec2:DescribeInstances",
"ec2:DescribeTags",
"ec2:DescribeNetworkInterfaces",
"ec2:DescribeInstanceTypes"
]
resources = ["*"]
}
statement {
sid = "CreateTags"
actions = ["ec2:CreateTags"]
resources = ["arn:${local.partition}:ec2:*:*:network-interface/*"]
}
}
# Note - we are keeping this to a minimum in hopes that its soon replaced with an AWS managed policy like `AmazonEKS_CNI_Policy`
resource "aws_iam_policy" "cni_ipv6_policy" {
count = var.create && var.create_cni_ipv6_iam_policy ? 1 : 0
# Will cause conflicts if trying to create on multiple clusters but necessary to reference by exact name in sub-modules
name = "AmazonEKS_CNI_IPv6_Policy"
description = "IAM policy for EKS CNI to assign IPV6 addresses"
policy = data.aws_iam_policy_document.cni_ipv6_policy[0].json
tags = var.tags
}
################################################################################
# Node Security Group
# Defaults follow https://docs.aws.amazon.com/eks/latest/userguide/sec-group-reqs.html
# Plus NTP/HTTPS (otherwise nodes fail to launch)
################################################################################
locals {
node_sg_name = coalesce(var.node_security_group_name, "${var.cluster_name}-node")
create_node_sg = var.create && var.create_node_security_group
node_security_group_id = local.create_node_sg ? aws_security_group.node[0].id : var.node_security_group_id
node_security_group_rules = {
ingress_cluster_443 = {
description = "Cluster API to node groups"
protocol = "tcp"
from_port = 443
to_port = 443
type = "ingress"
source_cluster_security_group = true
}
ingress_cluster_kubelet = {
description = "Cluster API to node kubelets"
protocol = "tcp"
from_port = 10250
to_port = 10250
type = "ingress"
source_cluster_security_group = true
}
ingress_self_coredns_tcp = {
description = "Node to node CoreDNS"
protocol = "tcp"
from_port = 53
to_port = 53
type = "ingress"
self = true
}
ingress_self_coredns_udp = {
description = "Node to node CoreDNS UDP"
protocol = "udp"
from_port = 53
to_port = 53
type = "ingress"
self = true
}
}
node_security_group_recommended_rules = { for k, v in {
ingress_nodes_ephemeral = {
description = "Node to node ingress on ephemeral ports"
protocol = "tcp"
from_port = 1025
to_port = 65535
type = "ingress"
self = true
}
# metrics-server
ingress_cluster_4443_webhook = {
description = "Cluster API to node 4443/tcp webhook"
protocol = "tcp"
from_port = 4443
to_port = 4443
type = "ingress"
source_cluster_security_group = true
}
# prometheus-adapter
ingress_cluster_6443_webhook = {
description = "Cluster API to node 6443/tcp webhook"
protocol = "tcp"
from_port = 6443
to_port = 6443
type = "ingress"
source_cluster_security_group = true
}
# Karpenter
ingress_cluster_8443_webhook = {
description = "Cluster API to node 8443/tcp webhook"
protocol = "tcp"
from_port = 8443
to_port = 8443
type = "ingress"
source_cluster_security_group = true
}
# ALB controller, NGINX
ingress_cluster_9443_webhook = {
description = "Cluster API to node 9443/tcp webhook"
protocol = "tcp"
from_port = 9443
to_port = 9443
type = "ingress"
source_cluster_security_group = true
}
egress_all = {
description = "Allow all egress"
protocol = "-1"
from_port = 0
to_port = 0
type = "egress"
cidr_blocks = ["0.0.0.0/0"]
ipv6_cidr_blocks = var.cluster_ip_family == "ipv6" ? ["::/0"] : null
}
} : k => v if var.node_security_group_enable_recommended_rules }
efa_security_group_rules = { for k, v in
{
ingress_all_self_efa = {
description = "Node to node EFA"
protocol = "-1"
from_port = 0
to_port = 0
type = "ingress"
self = true
}
egress_all_self_efa = {
description = "Node to node EFA"
protocol = "-1"
from_port = 0
to_port = 0
type = "egress"
self = true
}
} : k => v if var.enable_efa_support
}
}
resource "aws_security_group" "node" {
count = local.create_node_sg ? 1 : 0
name = var.node_security_group_use_name_prefix ? null : local.node_sg_name
name_prefix = var.node_security_group_use_name_prefix ? "${local.node_sg_name}${var.prefix_separator}" : null
description = var.node_security_group_description
vpc_id = var.vpc_id
tags = merge(
var.tags,
{
"Name" = local.node_sg_name
"kubernetes.io/cluster/${var.cluster_name}" = "owned"
},
var.node_security_group_tags
)
lifecycle {
create_before_destroy = true
}
}
resource "aws_security_group_rule" "node" {
for_each = { for k, v in merge(
local.efa_security_group_rules,
local.node_security_group_rules,
local.node_security_group_recommended_rules,
var.node_security_group_additional_rules,
) : k => v if local.create_node_sg }
# Required
security_group_id = aws_security_group.node[0].id
protocol = each.value.protocol
from_port = each.value.from_port
to_port = each.value.to_port
type = each.value.type
# Optional
description = lookup(each.value, "description", null)
cidr_blocks = lookup(each.value, "cidr_blocks", null)
ipv6_cidr_blocks = lookup(each.value, "ipv6_cidr_blocks", null)
prefix_list_ids = lookup(each.value, "prefix_list_ids", [])
self = lookup(each.value, "self", null)
source_security_group_id = try(each.value.source_cluster_security_group, false) ? local.cluster_security_group_id : lookup(each.value, "source_security_group_id", null)
}
################################################################################
# Fargate Profile
################################################################################
module "fargate_profile" {
source = "./modules/fargate-profile"
for_each = { for k, v in var.fargate_profiles : k => v if var.create && !local.create_outposts_local_cluster }
create = try(each.value.create, true)
# Fargate Profile
cluster_name = time_sleep.this[0].triggers["cluster_name"]
cluster_ip_family = var.cluster_ip_family
name = try(each.value.name, each.key)
subnet_ids = try(each.value.subnet_ids, var.fargate_profile_defaults.subnet_ids, var.subnet_ids)
selectors = try(each.value.selectors, var.fargate_profile_defaults.selectors, [])
timeouts = try(each.value.timeouts, var.fargate_profile_defaults.timeouts, {})
# IAM role
create_iam_role = try(each.value.create_iam_role, var.fargate_profile_defaults.create_iam_role, true)
iam_role_arn = try(each.value.iam_role_arn, var.fargate_profile_defaults.iam_role_arn, null)
iam_role_name = try(each.value.iam_role_name, var.fargate_profile_defaults.iam_role_name, null)
iam_role_use_name_prefix = try(each.value.iam_role_use_name_prefix, var.fargate_profile_defaults.iam_role_use_name_prefix, true)
iam_role_path = try(each.value.iam_role_path, var.fargate_profile_defaults.iam_role_path, null)
iam_role_description = try(each.value.iam_role_description, var.fargate_profile_defaults.iam_role_description, "Fargate profile IAM role")
iam_role_permissions_boundary = try(each.value.iam_role_permissions_boundary, var.fargate_profile_defaults.iam_role_permissions_boundary, null)
iam_role_tags = try(each.value.iam_role_tags, var.fargate_profile_defaults.iam_role_tags, {})
iam_role_attach_cni_policy = try(each.value.iam_role_attach_cni_policy, var.fargate_profile_defaults.iam_role_attach_cni_policy, true)
# To better understand why this `lookup()` logic is required, see:
# https://github.com/hashicorp/terraform/issues/31646#issuecomment-1217279031
iam_role_additional_policies = lookup(each.value, "iam_role_additional_policies", lookup(var.fargate_profile_defaults, "iam_role_additional_policies", {}))
create_iam_role_policy = try(each.value.create_iam_role_policy, var.fargate_profile_defaults.create_iam_role_policy, true)
iam_role_policy_statements = try(each.value.iam_role_policy_statements, var.fargate_profile_defaults.iam_role_policy_statements, [])
tags = merge(var.tags, try(each.value.tags, var.fargate_profile_defaults.tags, {}))
}
################################################################################
# EKS Managed Node Group
################################################################################
module "eks_managed_node_group" {
source = "./modules/eks-managed-node-group"
for_each = { for k, v in var.eks_managed_node_groups : k => v if var.create && !local.create_outposts_local_cluster }
create = try(each.value.create, true)
cluster_name = time_sleep.this[0].triggers["cluster_name"]
cluster_version = try(each.value.cluster_version, var.eks_managed_node_group_defaults.cluster_version, time_sleep.this[0].triggers["cluster_version"])
# EKS Managed Node Group
name = try(each.value.name, each.key)
use_name_prefix = try(each.value.use_name_prefix, var.eks_managed_node_group_defaults.use_name_prefix, true)
subnet_ids = try(each.value.subnet_ids, var.eks_managed_node_group_defaults.subnet_ids, var.subnet_ids)
min_size = try(each.value.min_size, var.eks_managed_node_group_defaults.min_size, 1)
max_size = try(each.value.max_size, var.eks_managed_node_group_defaults.max_size, 3)
desired_size = try(each.value.desired_size, var.eks_managed_node_group_defaults.desired_size, 1)
ami_id = try(each.value.ami_id, var.eks_managed_node_group_defaults.ami_id, "")
ami_type = try(each.value.ami_type, var.eks_managed_node_group_defaults.ami_type, null)
ami_release_version = try(each.value.ami_release_version, var.eks_managed_node_group_defaults.ami_release_version, null)
use_latest_ami_release_version = try(each.value.use_latest_ami_release_version, var.eks_managed_node_group_defaults.use_latest_ami_release_version, false)
capacity_type = try(each.value.capacity_type, var.eks_managed_node_group_defaults.capacity_type, null)
disk_size = try(each.value.disk_size, var.eks_managed_node_group_defaults.disk_size, null)
force_update_version = try(each.value.force_update_version, var.eks_managed_node_group_defaults.force_update_version, null)
instance_types = try(each.value.instance_types, var.eks_managed_node_group_defaults.instance_types, null)
labels = try(each.value.labels, var.eks_managed_node_group_defaults.labels, null)
remote_access = try(each.value.remote_access, var.eks_managed_node_group_defaults.remote_access, {})
taints = try(each.value.taints, var.eks_managed_node_group_defaults.taints, {})
update_config = try(each.value.update_config, var.eks_managed_node_group_defaults.update_config, local.default_update_config)
timeouts = try(each.value.timeouts, var.eks_managed_node_group_defaults.timeouts, {})
# User data
platform = try(each.value.platform, var.eks_managed_node_group_defaults.platform, "linux")
cluster_endpoint = try(time_sleep.this[0].triggers["cluster_endpoint"], "")
cluster_auth_base64 = try(time_sleep.this[0].triggers["cluster_certificate_authority_data"], "")
cluster_service_ipv4_cidr = var.cluster_service_ipv4_cidr
cluster_ip_family = var.cluster_ip_family
cluster_service_cidr = try(time_sleep.this[0].triggers["cluster_service_cidr"], "")
enable_bootstrap_user_data = try(each.value.enable_bootstrap_user_data, var.eks_managed_node_group_defaults.enable_bootstrap_user_data, false)
pre_bootstrap_user_data = try(each.value.pre_bootstrap_user_data, var.eks_managed_node_group_defaults.pre_bootstrap_user_data, "")
post_bootstrap_user_data = try(each.value.post_bootstrap_user_data, var.eks_managed_node_group_defaults.post_bootstrap_user_data, "")
bootstrap_extra_args = try(each.value.bootstrap_extra_args, var.eks_managed_node_group_defaults.bootstrap_extra_args, "")
user_data_template_path = try(each.value.user_data_template_path, var.eks_managed_node_group_defaults.user_data_template_path, "")
cloudinit_pre_nodeadm = try(each.value.cloudinit_pre_nodeadm, var.eks_managed_node_group_defaults.cloudinit_pre_nodeadm, [])
cloudinit_post_nodeadm = try(each.value.cloudinit_post_nodeadm, var.eks_managed_node_group_defaults.cloudinit_post_nodeadm, [])
# Launch Template
create_launch_template = try(each.value.create_launch_template, var.eks_managed_node_group_defaults.create_launch_template, true)
use_custom_launch_template = try(each.value.use_custom_launch_template, var.eks_managed_node_group_defaults.use_custom_launch_template, true)
launch_template_id = try(each.value.launch_template_id, var.eks_managed_node_group_defaults.launch_template_id, "")
launch_template_name = try(each.value.launch_template_name, var.eks_managed_node_group_defaults.launch_template_name, each.key)
launch_template_use_name_prefix = try(each.value.launch_template_use_name_prefix, var.eks_managed_node_group_defaults.launch_template_use_name_prefix, true)
launch_template_version = try(each.value.launch_template_version, var.eks_managed_node_group_defaults.launch_template_version, null)
launch_template_default_version = try(each.value.launch_template_default_version, var.eks_managed_node_group_defaults.launch_template_default_version, null)
update_launch_template_default_version = try(each.value.update_launch_template_default_version, var.eks_managed_node_group_defaults.update_launch_template_default_version, true)
launch_template_description = try(each.value.launch_template_description, var.eks_managed_node_group_defaults.launch_template_description, "Custom launch template for ${try(each.value.name, each.key)} EKS managed node group")
launch_template_tags = try(each.value.launch_template_tags, var.eks_managed_node_group_defaults.launch_template_tags, {})
tag_specifications = try(each.value.tag_specifications, var.eks_managed_node_group_defaults.tag_specifications, ["instance", "volume", "network-interface"])
ebs_optimized = try(each.value.ebs_optimized, var.eks_managed_node_group_defaults.ebs_optimized, null)
key_name = try(each.value.key_name, var.eks_managed_node_group_defaults.key_name, null)
disable_api_termination = try(each.value.disable_api_termination, var.eks_managed_node_group_defaults.disable_api_termination, null)
kernel_id = try(each.value.kernel_id, var.eks_managed_node_group_defaults.kernel_id, null)
ram_disk_id = try(each.value.ram_disk_id, var.eks_managed_node_group_defaults.ram_disk_id, null)
block_device_mappings = try(each.value.block_device_mappings, var.eks_managed_node_group_defaults.block_device_mappings, {})
capacity_reservation_specification = try(each.value.capacity_reservation_specification, var.eks_managed_node_group_defaults.capacity_reservation_specification, {})
cpu_options = try(each.value.cpu_options, var.eks_managed_node_group_defaults.cpu_options, {})
credit_specification = try(each.value.credit_specification, var.eks_managed_node_group_defaults.credit_specification, {})
elastic_gpu_specifications = try(each.value.elastic_gpu_specifications, var.eks_managed_node_group_defaults.elastic_gpu_specifications, {})
elastic_inference_accelerator = try(each.value.elastic_inference_accelerator, var.eks_managed_node_group_defaults.elastic_inference_accelerator, {})
enclave_options = try(each.value.enclave_options, var.eks_managed_node_group_defaults.enclave_options, {})
instance_market_options = try(each.value.instance_market_options, var.eks_managed_node_group_defaults.instance_market_options, {})
license_specifications = try(each.value.license_specifications, var.eks_managed_node_group_defaults.license_specifications, {})
metadata_options = try(each.value.metadata_options, var.eks_managed_node_group_defaults.metadata_options, local.metadata_options)
enable_monitoring = try(each.value.enable_monitoring, var.eks_managed_node_group_defaults.enable_monitoring, true)
enable_efa_support = try(each.value.enable_efa_support, var.eks_managed_node_group_defaults.enable_efa_support, false)
create_placement_group = try(each.value.create_placement_group, var.eks_managed_node_group_defaults.create_placement_group, false)
placement = try(each.value.placement, var.eks_managed_node_group_defaults.placement, {})
placement_group_az = try(each.value.placement_group_az, var.eks_managed_node_group_defaults.placement_group_az, null)
placement_group_strategy = try(each.value.placement_group_strategy, var.eks_managed_node_group_defaults.placement_group_strategy, "cluster")
network_interfaces = try(each.value.network_interfaces, var.eks_managed_node_group_defaults.network_interfaces, [])
maintenance_options = try(each.value.maintenance_options, var.eks_managed_node_group_defaults.maintenance_options, {})
private_dns_name_options = try(each.value.private_dns_name_options, var.eks_managed_node_group_defaults.private_dns_name_options, {})
# IAM role
create_iam_role = try(each.value.create_iam_role, var.eks_managed_node_group_defaults.create_iam_role, true)
iam_role_arn = try(each.value.iam_role_arn, var.eks_managed_node_group_defaults.iam_role_arn, null)
iam_role_name = try(each.value.iam_role_name, var.eks_managed_node_group_defaults.iam_role_name, null)
iam_role_use_name_prefix = try(each.value.iam_role_use_name_prefix, var.eks_managed_node_group_defaults.iam_role_use_name_prefix, true)
iam_role_path = try(each.value.iam_role_path, var.eks_managed_node_group_defaults.iam_role_path, null)
iam_role_description = try(each.value.iam_role_description, var.eks_managed_node_group_defaults.iam_role_description, "EKS managed node group IAM role")
iam_role_permissions_boundary = try(each.value.iam_role_permissions_boundary, var.eks_managed_node_group_defaults.iam_role_permissions_boundary, null)
iam_role_tags = try(each.value.iam_role_tags, var.eks_managed_node_group_defaults.iam_role_tags, {})
iam_role_attach_cni_policy = try(each.value.iam_role_attach_cni_policy, var.eks_managed_node_group_defaults.iam_role_attach_cni_policy, true)
# To better understand why this `lookup()` logic is required, see:
# https://github.com/hashicorp/terraform/issues/31646#issuecomment-1217279031
iam_role_additional_policies = lookup(each.value, "iam_role_additional_policies", lookup(var.eks_managed_node_group_defaults, "iam_role_additional_policies", {}))
create_iam_role_policy = try(each.value.create_iam_role_policy, var.eks_managed_node_group_defaults.create_iam_role_policy, true)
iam_role_policy_statements = try(each.value.iam_role_policy_statements, var.eks_managed_node_group_defaults.iam_role_policy_statements, [])
# Autoscaling group schedule
create_schedule = try(each.value.create_schedule, var.eks_managed_node_group_defaults.create_schedule, true)
schedules = try(each.value.schedules, var.eks_managed_node_group_defaults.schedules, {})
# Security group
vpc_security_group_ids = compact(concat([local.node_security_group_id], try(each.value.vpc_security_group_ids, var.eks_managed_node_group_defaults.vpc_security_group_ids, [])))
cluster_primary_security_group_id = try(each.value.attach_cluster_primary_security_group, var.eks_managed_node_group_defaults.attach_cluster_primary_security_group, false) ? aws_eks_cluster.this[0].vpc_config[0].cluster_security_group_id : null
tags = merge(var.tags, try(each.value.tags, var.eks_managed_node_group_defaults.tags, {}))
}
################################################################################
# Self Managed Node Group
################################################################################
module "self_managed_node_group" {
source = "./modules/self-managed-node-group"
for_each = { for k, v in var.self_managed_node_groups : k => v if var.create }
create = try(each.value.create, true)
cluster_name = time_sleep.this[0].triggers["cluster_name"]
# Autoscaling Group
create_autoscaling_group = try(each.value.create_autoscaling_group, var.self_managed_node_group_defaults.create_autoscaling_group, true)
name = try(each.value.name, each.key)
use_name_prefix = try(each.value.use_name_prefix, var.self_managed_node_group_defaults.use_name_prefix, true)
availability_zones = try(each.value.availability_zones, var.self_managed_node_group_defaults.availability_zones, null)
subnet_ids = try(each.value.subnet_ids, var.self_managed_node_group_defaults.subnet_ids, var.subnet_ids)
min_size = try(each.value.min_size, var.self_managed_node_group_defaults.min_size, 0)
max_size = try(each.value.max_size, var.self_managed_node_group_defaults.max_size, 3)
desired_size = try(each.value.desired_size, var.self_managed_node_group_defaults.desired_size, 1)
capacity_rebalance = try(each.value.capacity_rebalance, var.self_managed_node_group_defaults.capacity_rebalance, null)
min_elb_capacity = try(each.value.min_elb_capacity, var.self_managed_node_group_defaults.min_elb_capacity, null)
wait_for_elb_capacity = try(each.value.wait_for_elb_capacity, var.self_managed_node_group_defaults.wait_for_elb_capacity, null)
wait_for_capacity_timeout = try(each.value.wait_for_capacity_timeout, var.self_managed_node_group_defaults.wait_for_capacity_timeout, null)
default_cooldown = try(each.value.default_cooldown, var.self_managed_node_group_defaults.default_cooldown, null)
default_instance_warmup = try(each.value.default_instance_warmup, var.self_managed_node_group_defaults.default_instance_warmup, null)
protect_from_scale_in = try(each.value.protect_from_scale_in, var.self_managed_node_group_defaults.protect_from_scale_in, null)
context = try(each.value.context, var.self_managed_node_group_defaults.context, null)
target_group_arns = try(each.value.target_group_arns, var.self_managed_node_group_defaults.target_group_arns, [])
create_placement_group = try(each.value.create_placement_group, var.self_managed_node_group_defaults.create_placement_group, false)
placement_group = try(each.value.placement_group, var.self_managed_node_group_defaults.placement_group, null)
placement_group_az = try(each.value.placement_group_az, var.self_managed_node_group_defaults.placement_group_az, null)
health_check_type = try(each.value.health_check_type, var.self_managed_node_group_defaults.health_check_type, null)
health_check_grace_period = try(each.value.health_check_grace_period, var.self_managed_node_group_defaults.health_check_grace_period, null)
ignore_failed_scaling_activities = try(each.value.ignore_failed_scaling_activities, var.self_managed_node_group_defaults.ignore_failed_scaling_activities, null)
force_delete = try(each.value.force_delete, var.self_managed_node_group_defaults.force_delete, null)
force_delete_warm_pool = try(each.value.force_delete_warm_pool, var.self_managed_node_group_defaults.force_delete_warm_pool, null)
termination_policies = try(each.value.termination_policies, var.self_managed_node_group_defaults.termination_policies, [])
suspended_processes = try(each.value.suspended_processes, var.self_managed_node_group_defaults.suspended_processes, [])
max_instance_lifetime = try(each.value.max_instance_lifetime, var.self_managed_node_group_defaults.max_instance_lifetime, null)
enabled_metrics = try(each.value.enabled_metrics, var.self_managed_node_group_defaults.enabled_metrics, [])
metrics_granularity = try(each.value.metrics_granularity, var.self_managed_node_group_defaults.metrics_granularity, null)
service_linked_role_arn = try(each.value.service_linked_role_arn, var.self_managed_node_group_defaults.service_linked_role_arn, null)
initial_lifecycle_hooks = try(each.value.initial_lifecycle_hooks, var.self_managed_node_group_defaults.initial_lifecycle_hooks, [])
instance_maintenance_policy = try(each.value.instance_maintenance_policy, var.self_managed_node_group_defaults.instance_maintenance_policy, {})
instance_refresh = try(each.value.instance_refresh, var.self_managed_node_group_defaults.instance_refresh, local.default_instance_refresh)
use_mixed_instances_policy = try(each.value.use_mixed_instances_policy, var.self_managed_node_group_defaults.use_mixed_instances_policy, false)
mixed_instances_policy = try(each.value.mixed_instances_policy, var.self_managed_node_group_defaults.mixed_instances_policy, null)
warm_pool = try(each.value.warm_pool, var.self_managed_node_group_defaults.warm_pool, {})
delete_timeout = try(each.value.delete_timeout, var.self_managed_node_group_defaults.delete_timeout, null)
autoscaling_group_tags = try(each.value.autoscaling_group_tags, var.self_managed_node_group_defaults.autoscaling_group_tags, {})
# User data
platform = try(each.value.platform, var.self_managed_node_group_defaults.platform, null)
# TODO - update this when `var.platform` is removed in v21.0
ami_type = try(each.value.ami_type, var.self_managed_node_group_defaults.ami_type, "AL2_x86_64")
cluster_endpoint = try(time_sleep.this[0].triggers["cluster_endpoint"], "")
cluster_auth_base64 = try(time_sleep.this[0].triggers["cluster_certificate_authority_data"], "")
cluster_service_cidr = try(time_sleep.this[0].triggers["cluster_service_cidr"], "")
cluster_ip_family = var.cluster_ip_family
pre_bootstrap_user_data = try(each.value.pre_bootstrap_user_data, var.self_managed_node_group_defaults.pre_bootstrap_user_data, "")
post_bootstrap_user_data = try(each.value.post_bootstrap_user_data, var.self_managed_node_group_defaults.post_bootstrap_user_data, "")
bootstrap_extra_args = try(each.value.bootstrap_extra_args, var.self_managed_node_group_defaults.bootstrap_extra_args, "")
user_data_template_path = try(each.value.user_data_template_path, var.self_managed_node_group_defaults.user_data_template_path, "")
cloudinit_pre_nodeadm = try(each.value.cloudinit_pre_nodeadm, var.self_managed_node_group_defaults.cloudinit_pre_nodeadm, [])
cloudinit_post_nodeadm = try(each.value.cloudinit_post_nodeadm, var.self_managed_node_group_defaults.cloudinit_post_nodeadm, [])
# Launch Template
create_launch_template = try(each.value.create_launch_template, var.self_managed_node_group_defaults.create_launch_template, true)
launch_template_id = try(each.value.launch_template_id, var.self_managed_node_group_defaults.launch_template_id, "")
launch_template_name = try(each.value.launch_template_name, var.self_managed_node_group_defaults.launch_template_name, each.key)
launch_template_use_name_prefix = try(each.value.launch_template_use_name_prefix, var.self_managed_node_group_defaults.launch_template_use_name_prefix, true)
launch_template_version = try(each.value.launch_template_version, var.self_managed_node_group_defaults.launch_template_version, null)
launch_template_default_version = try(each.value.launch_template_default_version, var.self_managed_node_group_defaults.launch_template_default_version, null)
update_launch_template_default_version = try(each.value.update_launch_template_default_version, var.self_managed_node_group_defaults.update_launch_template_default_version, true)
launch_template_description = try(each.value.launch_template_description, var.self_managed_node_group_defaults.launch_template_description, "Custom launch template for ${try(each.value.name, each.key)} self managed node group")
launch_template_tags = try(each.value.launch_template_tags, var.self_managed_node_group_defaults.launch_template_tags, {})
tag_specifications = try(each.value.tag_specifications, var.self_managed_node_group_defaults.tag_specifications, ["instance", "volume", "network-interface"])
ebs_optimized = try(each.value.ebs_optimized, var.self_managed_node_group_defaults.ebs_optimized, null)
ami_id = try(each.value.ami_id, var.self_managed_node_group_defaults.ami_id, "")
cluster_version = try(each.value.cluster_version, var.self_managed_node_group_defaults.cluster_version, time_sleep.this[0].triggers["cluster_version"])
instance_type = try(each.value.instance_type, var.self_managed_node_group_defaults.instance_type, "m6i.large")
key_name = try(each.value.key_name, var.self_managed_node_group_defaults.key_name, null)
disable_api_termination = try(each.value.disable_api_termination, var.self_managed_node_group_defaults.disable_api_termination, null)
instance_initiated_shutdown_behavior = try(each.value.instance_initiated_shutdown_behavior, var.self_managed_node_group_defaults.instance_initiated_shutdown_behavior, null)
kernel_id = try(each.value.kernel_id, var.self_managed_node_group_defaults.kernel_id, null)
ram_disk_id = try(each.value.ram_disk_id, var.self_managed_node_group_defaults.ram_disk_id, null)
block_device_mappings = try(each.value.block_device_mappings, var.self_managed_node_group_defaults.block_device_mappings, {})
capacity_reservation_specification = try(each.value.capacity_reservation_specification, var.self_managed_node_group_defaults.capacity_reservation_specification, {})
cpu_options = try(each.value.cpu_options, var.self_managed_node_group_defaults.cpu_options, {})
credit_specification = try(each.value.credit_specification, var.self_managed_node_group_defaults.credit_specification, {})
elastic_gpu_specifications = try(each.value.elastic_gpu_specifications, var.self_managed_node_group_defaults.elastic_gpu_specifications, {})
elastic_inference_accelerator = try(each.value.elastic_inference_accelerator, var.self_managed_node_group_defaults.elastic_inference_accelerator, {})
enclave_options = try(each.value.enclave_options, var.self_managed_node_group_defaults.enclave_options, {})
hibernation_options = try(each.value.hibernation_options, var.self_managed_node_group_defaults.hibernation_options, {})
instance_requirements = try(each.value.instance_requirements, var.self_managed_node_group_defaults.instance_requirements, {})
instance_market_options = try(each.value.instance_market_options, var.self_managed_node_group_defaults.instance_market_options, {})
license_specifications = try(each.value.license_specifications, var.self_managed_node_group_defaults.license_specifications, {})
metadata_options = try(each.value.metadata_options, var.self_managed_node_group_defaults.metadata_options, local.metadata_options)
enable_monitoring = try(each.value.enable_monitoring, var.self_managed_node_group_defaults.enable_monitoring, true)
enable_efa_support = try(each.value.enable_efa_support, var.self_managed_node_group_defaults.enable_efa_support, false)
network_interfaces = try(each.value.network_interfaces, var.self_managed_node_group_defaults.network_interfaces, [])
placement = try(each.value.placement, var.self_managed_node_group_defaults.placement, {})
maintenance_options = try(each.value.maintenance_options, var.self_managed_node_group_defaults.maintenance_options, {})
private_dns_name_options = try(each.value.private_dns_name_options, var.self_managed_node_group_defaults.private_dns_name_options, {})
# IAM role
create_iam_instance_profile = try(each.value.create_iam_instance_profile, var.self_managed_node_group_defaults.create_iam_instance_profile, true)
iam_instance_profile_arn = try(each.value.iam_instance_profile_arn, var.self_managed_node_group_defaults.iam_instance_profile_arn, null)
iam_role_name = try(each.value.iam_role_name, var.self_managed_node_group_defaults.iam_role_name, null)
iam_role_use_name_prefix = try(each.value.iam_role_use_name_prefix, var.self_managed_node_group_defaults.iam_role_use_name_prefix, true)
iam_role_path = try(each.value.iam_role_path, var.self_managed_node_group_defaults.iam_role_path, null)
iam_role_description = try(each.value.iam_role_description, var.self_managed_node_group_defaults.iam_role_description, "Self managed node group IAM role")
iam_role_permissions_boundary = try(each.value.iam_role_permissions_boundary, var.self_managed_node_group_defaults.iam_role_permissions_boundary, null)
iam_role_tags = try(each.value.iam_role_tags, var.self_managed_node_group_defaults.iam_role_tags, {})
iam_role_attach_cni_policy = try(each.value.iam_role_attach_cni_policy, var.self_managed_node_group_defaults.iam_role_attach_cni_policy, true)
# To better understand why this `lookup()` logic is required, see:
# https://github.com/hashicorp/terraform/issues/31646#issuecomment-1217279031
iam_role_additional_policies = lookup(each.value, "iam_role_additional_policies", lookup(var.self_managed_node_group_defaults, "iam_role_additional_policies", {}))
create_iam_role_policy = try(each.value.create_iam_role_policy, var.self_managed_node_group_defaults.create_iam_role_policy, true)
iam_role_policy_statements = try(each.value.iam_role_policy_statements, var.self_managed_node_group_defaults.iam_role_policy_statements, [])
# Access entry
create_access_entry = try(each.value.create_access_entry, var.self_managed_node_group_defaults.create_access_entry, true)
iam_role_arn = try(each.value.iam_role_arn, var.self_managed_node_group_defaults.iam_role_arn, null)
# Autoscaling group schedule
create_schedule = try(each.value.create_schedule, var.self_managed_node_group_defaults.create_schedule, true)
schedules = try(each.value.schedules, var.self_managed_node_group_defaults.schedules, {})
# Security group
vpc_security_group_ids = compact(concat([local.node_security_group_id], try(each.value.vpc_security_group_ids, var.self_managed_node_group_defaults.vpc_security_group_ids, [])))
cluster_primary_security_group_id = try(each.value.attach_cluster_primary_security_group, var.self_managed_node_group_defaults.attach_cluster_primary_security_group, false) ? aws_eks_cluster.this[0].vpc_config[0].cluster_security_group_id : null
tags = merge(var.tags, try(each.value.tags, var.self_managed_node_group_defaults.tags, {}))
}