forked from aws-samples/sagemaker-studio-mlflow-integration
-
Notifications
You must be signed in to change notification settings - Fork 0
/
mlflow-vpc-stack.ts
421 lines (377 loc) · 13 KB
/
mlflow-vpc-stack.ts
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
import * as cdk from 'aws-cdk-lib';
import { Construct } from 'constructs';
import * as elbv2 from "aws-cdk-lib/aws-elasticloadbalancingv2";
import * as ec2 from "aws-cdk-lib/aws-ec2";
import * as ecs from "aws-cdk-lib/aws-ecs";
import * as iam from "aws-cdk-lib/aws-iam";
import * as logs from "aws-cdk-lib/aws-logs";
import * as servicediscovery from "aws-cdk-lib/aws-servicediscovery";
import * as secretsmanager from 'aws-cdk-lib/aws-secretsmanager';
import * as s3 from 'aws-cdk-lib/aws-s3';
import { CfnDBCluster, CfnDBSubnetGroup } from 'aws-cdk-lib/aws-rds';
import { NagSuppressions } from 'cdk-nag'
const { Protocol } = elbv2;
const dbName = "mlflowdb"
const dbPort = 5432
const dbUsername = "master"
const clusterName = "mlflowCluster"
const serviceName = "mlflowService"
const cidr = "10.0.0.0/16"
const containerPort = 5000
export class MLflowVpcStack extends cdk.Stack {
// Export Vpc, ALB Listener, and Mlflow secret ARN
public readonly httpApiListener: elbv2.NetworkListener;
public readonly vpc: ec2.Vpc;
public readonly httpApiInternalNLB: elbv2.NetworkLoadBalancer;
public readonly accessLogs: s3.Bucket;
readonly bucketName = `mlflow-${this.account}-${this.region}`
readonly accesslogBucketName = `accesslogs-${this.account}-${this.region}`
constructor(
scope: Construct,
id: string,
props?: cdk.StackProps
) {
super(scope, id, props);
const logGroup = new logs.LogGroup(this, 'MyVpcLogGroup');
const flowLogsRole = new iam.Role(this, 'flowLogsRole', {
assumedBy: new iam.ServicePrincipal('vpc-flow-logs.amazonaws.com')
});
// VPC
this.vpc = new ec2.Vpc(this, 'MLFlowVPC', {
ipAddresses: ec2.IpAddresses.cidr(cidr),
natGateways: 1,
maxAzs: 2,
subnetConfiguration: [
{
name: 'public',
subnetType: ec2.SubnetType.PUBLIC,
cidrMask: 24,
mapPublicIpOnLaunch: false
},
{
name: 'private',
subnetType: ec2.SubnetType.PRIVATE_WITH_EGRESS,
cidrMask: 26,
},
{
name: 'isolated',
subnetType: ec2.SubnetType.PRIVATE_ISOLATED,
cidrMask: 28,
},
],
});
new ec2.FlowLog(this, 'FlowLog', {
resourceType: ec2.FlowLogResourceType.fromVpc(this.vpc),
destination: ec2.FlowLogDestination.toCloudWatchLogs(logGroup, flowLogsRole)
});
this.accessLogs = new s3.Bucket(this, "accessLogs", {
versioned: false,
bucketName: this.accesslogBucketName,
publicReadAccess: false,
blockPublicAccess: s3.BlockPublicAccess.BLOCK_ALL,
removalPolicy: cdk.RemovalPolicy.DESTROY,
autoDeleteObjects: true,
encryption: s3.BucketEncryption.KMS_MANAGED,
enforceSSL: true,
objectOwnership: s3.ObjectOwnership.BUCKET_OWNER_PREFERRED
})
// mlflow S3 bucket
const mlFlowBucket = new s3.Bucket(this, "mlFlowBucket", {
versioned: false,
bucketName: this.bucketName,
publicReadAccess: false,
blockPublicAccess: s3.BlockPublicAccess.BLOCK_ALL,
removalPolicy: cdk.RemovalPolicy.DESTROY,
autoDeleteObjects: true,
encryption: s3.BucketEncryption.KMS_MANAGED,
enforceSSL: true,
serverAccessLogsBucket: this.accessLogs,
serverAccessLogsPrefix: 'mlflow-server'
})
// DB SubnetGroup
const subnetIds: string[] = [];
this.vpc.isolatedSubnets.forEach((subnet, index) => {
subnetIds.push(subnet.subnetId);
});
const dbSubnetGroup: CfnDBSubnetGroup = new CfnDBSubnetGroup(this, 'AuroraSubnetGroup', {
dbSubnetGroupDescription: 'Subnet group to access aurora',
dbSubnetGroupName: 'aurora-serverless-subnet-group',
subnetIds
});
// DB Credentials
const databaseCredentialsSecret = new secretsmanager.Secret(this, 'DBCredentialsSecret', {
secretName: `mlflow-database-credentials`,
generateSecretString: {
secretStringTemplate: JSON.stringify({
username: dbUsername,
}),
excludePunctuation: true,
includeSpace: false,
generateStringKey: 'password'
}
});
// DB SecurityGroup
const dbClusterSecurityGroup = new ec2.SecurityGroup(this, 'DBClusterSecurityGroup',
{
vpc: this.vpc,
allowAllOutbound: false
}
);
dbClusterSecurityGroup.addIngressRule(ec2.Peer.ipv4(cidr), ec2.Port.tcp(dbPort));
const dbConfig = {
dbClusterIdentifier: `${serviceName}-cluster`,
engineMode: 'serverless',
engine: 'aurora-postgresql',
engineVersion: '11.16',
databaseName: dbName,
deletionProtection: false,
masterUsername: databaseCredentialsSecret.secretValueFromJson('username').toString(),
masterUserPassword: databaseCredentialsSecret.secretValueFromJson('password').toString(),
dbSubnetGroupName: dbSubnetGroup.dbSubnetGroupName,
scalingConfiguration: {
autoPause: true,
maxCapacity: 2,
minCapacity: 2,
secondsUntilAutoPause: 3600,
},
vpcSecurityGroupIds: [
dbClusterSecurityGroup.securityGroupId
],
storageEncrypted: true,
removalPolicy: cdk.RemovalPolicy.DESTROY // Delete everything
};
// 👇 RDS Cluster
const rdsCluster = new CfnDBCluster(this, 'DBCluster', dbConfig);
rdsCluster.addDependency(dbSubnetGroup)
// 👇 ECS Cluster
const cluster = new ecs.Cluster(this, "Fargate Cluster", {
vpc: this.vpc,
clusterName: clusterName,
containerInsights: true
});
// 👇 Cloud Map Namespace
const dnsNamespace = new servicediscovery.PrivateDnsNamespace(
this,
"DnsNamespace",
{
name: "http-api.local",
vpc: this.vpc,
description: "Private DnsNamespace for Microservices",
}
);
const withoutPolicyUpdatesOptions: iam.WithoutPolicyUpdatesOptions = {
addGrantsToResources: false,
};
// 👇 Fargate Task Role
const taskrole = new iam.Role(this, "ecsTaskExecutionRole", {
assumedBy: new iam.ServicePrincipal("ecs-tasks.amazonaws.com"),
managedPolicies: [
iam.ManagedPolicy.fromAwsManagedPolicyName("service-role/AmazonECSTaskExecutionRolePolicy")
],
inlinePolicies: {
s3Bucket: new iam.PolicyDocument({
statements:[
new iam.PolicyStatement({
effect: iam.Effect.ALLOW,
resources: [
`arn:aws:s3:::${this.bucketName}`,
`arn:aws:s3:::${this.bucketName}/*`
],
actions: [
"s3:ListBucket",
"s3:GetObject",
"s3:PutObject",
"s3:DeleteObject",
"s3:PutObjectTagging",
"s3:DeleteObjectTagging",
"s3:GetBucketTagging",
"s3:GetObjectTagging"
]
})
]
}),
secretsManagerRestricted: new iam.PolicyDocument({
statements: [
new iam.PolicyStatement({
effect: iam.Effect.ALLOW,
resources: [
databaseCredentialsSecret.secretArn
],
actions: [
"secretsmanager:GetResourcePolicy",
"secretsmanager:GetSecretValue",
"secretsmanager:DescribeSecret",
"secretsmanager:ListSecretVersionIds"
]
}),
]
})
}
});
// 👇 Task Definitions
const mlflowTaskDefinition = new ecs.FargateTaskDefinition(
this,
"mlflowTaskDef",
{
taskRole: taskrole.withoutPolicyUpdates(withoutPolicyUpdatesOptions),
executionRole: taskrole.withoutPolicyUpdates(withoutPolicyUpdatesOptions),
family: "mlFlowStack",
cpu: 512,
memoryLimitMiB: 1024
},
);
// 👇 Log Groups
const mlflowServiceLogGroup = new logs.LogGroup(this, "mlflowServiceLogGroup", {
logGroupName: "/ecs/mlflowService",
removalPolicy: cdk.RemovalPolicy.DESTROY,
});
const mlflowServiceLogDriver = new ecs.AwsLogDriver({
logGroup: mlflowServiceLogGroup,
streamPrefix: "mlflowService",
});
// 👇 MlFlow Task Container
const mlflowServiceContainer = mlflowTaskDefinition.addContainer(
"mlflowContainer",
{
containerName: "mlflowContainer",
essential: true,
memoryReservationMiB: 1024,
cpu: 512,
portMappings: [{
containerPort: containerPort,
protocol: ecs.Protocol.TCP,
}],
image: ecs.ContainerImage.fromAsset('../src/mlflow', {}),
environment: {
'BUCKET': `s3://${mlFlowBucket.bucketName}`,
'HOST': rdsCluster.attrEndpointAddress,
'PORT': `${dbPort}`,
'DATABASE': dbName
},
secrets: {
USERNAME: ecs.Secret.fromSecretsManager(databaseCredentialsSecret, 'username'),
PASSWORD: ecs.Secret.fromSecretsManager(databaseCredentialsSecret, 'password')
},
logging: mlflowServiceLogDriver,
});
NagSuppressions.addResourceSuppressions(mlflowServiceContainer, [
{
id: 'AwsSolutions-ECS2',
reason: 'ENV variables passed do not contain secrets'
},
])
// Security Group
const mlflowServiceSecGrp = new ec2.SecurityGroup(
this,
"mlflowServiceSecurityGroup",
{
vpc: this.vpc,
}
);
mlflowServiceSecGrp.addIngressRule(ec2.Peer.ipv4(cidr), ec2.Port.tcp(containerPort), 'Allow internal access to the container port');
mlflowServiceSecGrp.addIngressRule(ec2.Peer.ipv4(cidr), ec2.Port.tcp(80), 'Allow internal access to the container port');
// 👇 Fargate Services
const mlflowService = new ecs.FargateService(this, "mlflowService", {
cluster: cluster,
serviceName: serviceName,
taskDefinition: mlflowTaskDefinition,
assignPublicIp: false,
desiredCount: 2,
securityGroups: [mlflowServiceSecGrp],
cloudMapOptions: {
name: "mlflowService",
cloudMapNamespace: dnsNamespace,
},
});
// 👇 NLB
this.httpApiInternalNLB = new elbv2.NetworkLoadBalancer(
this,
"httpapiInternalALB",
{
vpc: this.vpc,
internetFacing: false,
}
);
// NLB Listener
this.httpApiListener = this.httpApiInternalNLB.addListener("httpapiListener", {
port: 80,
protocol: Protocol.TCP
});
// 👇 Target Groups
const mlflowServiceTargetGroup = this.httpApiListener.addTargets(
"mlflowServiceTargetGroup",
{
targets: [
mlflowService.loadBalancerTarget(
{
containerName: 'mlflowContainer',
containerPort: 5000
}
)
],
port: 80,
}
);
// 👇 Task Auto Scaling
const autoScaling = mlflowService.autoScaleTaskCount({ maxCapacity: 6 });
autoScaling.scaleOnCpuUtilization('CpuScaling', {
targetUtilizationPercent: 70,
scaleInCooldown: cdk.Duration.seconds(60),
scaleOutCooldown: cdk.Duration.seconds(60),
});
NagSuppressions.addResourceSuppressions(mlflowTaskDefinition, [
{
id: 'AwsSolutions-ECS2',
reason: 'ENV variables passed do not contain secrets'
},
])
NagSuppressions.addResourceSuppressions(taskrole, [
{
id: 'AwsSolutions-IAM5',
reason: 'The task owns this bucket and it should have full permissions on the objects',
appliesTo: [`Resource::arn:aws:s3:::${this.bucketName}/*`]
},
{
id: 'AwsSolutions-IAM4',
reason: 'The task needs access to this managed policy',
appliesTo: ['Policy::arn:<AWS::Partition>:iam::aws:policy/service-role/AmazonECSTaskExecutionRolePolicy']
}
]
)
NagSuppressions.addResourceSuppressions(databaseCredentialsSecret, [
{
id: 'AwsSolutions-SMG4',
reason: 'MLflow does not support database credentials rotation'
}
])
NagSuppressions.addResourceSuppressions(this.accessLogs, [
{
id: 'AwsSolutions-S1',
reason: 'This is a already an access log bucket'
}
])
NagSuppressions.addResourceSuppressions(rdsCluster, [
{
id: 'AwsSolutions-RDS11',
reason: 'We want to avoid creating confusion by obfuscating the standard Postgres port'
},
{
id: 'AwsSolutions-RDS10',
reason: 'This is a sample and we encourage users to clean up after trying the solution'
},
{
id: 'AwsSolutions-RDS6',
reason: 'MLflow does not support IAM authentication for the DB layer'
}
])
NagSuppressions.addResourceSuppressions(this.httpApiInternalNLB, [
{
id: 'AwsSolutions-ELB2',
reason: 'This is an internal-only NLB listening on port 80. Access logs for NLB only works for a TLS listener as per documentation in https://docs.aws.amazon.com/elasticloadbalancing/latest/network/load-balancer-access-logs.html'
}]
)
new cdk.CfnOutput(this, "ALB Dns Name : ", {
value: this.httpApiInternalNLB.loadBalancerDnsName,
});
}
}