anyone know how I can get pulumi to ignore cloudwatch resources if it already exists?

* creating CloudWatch Logs Log Group (user-redis-engine_log-eks-dev-20231220): operation error CloudWatch Logs: CreateLogGroup, https response error StatusCode: 400, RequestID: b943f5bf-7285-4a17-8687-e5cccda992ba, ResourceAlreadyExistsException: The specified log group already exists

export const userSlowLog = new aws.cloudwatch.LogGroup(userRedisSlowLogName, {
  name: userRedisSlowLogName,
  tags: additionalDefaultTags,
}, { ignoreChanges: ["name"] });

export const userRedisEngineLog = new aws.cloudwatch.LogGroup(userRedisEngineLogName, {
  name: userRedisEngineLogName,
  tags: additionalDefaultTags,
}, { ignoreChanges: ["name"] });

That's an AWS error, not a Pulumi one. You can't create a thing that already exists. The best option is to import it into Pulumi.

that is challenging as I wanted to rerun this stack multiple times.

That's not challenging. Importing is easy and you only do it once.

my experience is the opposite of that but I'll try

We're here to help! You just need to add

import: "user-redis-engine_log-eks-dev-20231220"

after

ignoreChanges: ["name"]

and it'll import.

that was an attempt to bypass the stoppage to pulumi up. I've given up on that and am just trying this, on a new run with

export const userSlowLog = new aws.cloudwatch.LogGroup(userRedisSlowLogName, {
  name: userRedisSlowLogName,
  tags: additionalDefaultTags,
});

export const userRedisEngineLog = new aws.cloudwatch.LogGroup(userRedisEngineLogName, {
  name: userRedisEngineLogName,
  tags: additionalDefaultTags,
});

Is there a reason to use

name

? It's generally recommended not to.

no particular reason happy to remove, one sec for the stack.

The

name

property should be used only when you need a "well-known" name: maybe for legacy reasons, or because a 3rd party unconfigurable app requires a resource to have a specific name.

I am trying to create this in a stack that also creates eks and a vpc and if it is possible use the same vpc for both resources (I think redis needs to make its own subnets in it and I think that is fine):

const userRedisSlowLogName = `user-redis-slow_log-${stack}`;
const userRedisEngineLogName = `user-redis-engine_log-${stack}`;

export const userSlowLog = new aws.cloudwatch.LogGroup(userRedisSlowLogName, {
  name: userRedisSlowLogName,
  tags: additionalDefaultTags,
});

export const userRedisEngineLog = new aws.cloudwatch.LogGroup(userRedisEngineLogName, {
  name: userRedisEngineLogName,
  tags: additionalDefaultTags,
});

const redisName = `user-redis-${stack}`;
const redisSubnetGroupName = `user-redis-subnet-group-${stack}`;
const redisEngineVersion = '7.1'; // Use the latest stable version
const redisMaintenanceWindow = 'sun:03:00-sun:04:00'; // Set to off-peak hours
const redisNodeType = 'cache.r6g.large'; // Use a memory-optimized instance type
const redisNumCacheClusters = 3; // Increase number of nodes for high availability and read scalability

export const userRedisSubnetGroup = new aws.elasticache.SubnetGroup(redisSubnetGroupName, {
  description: 'Managed by Pulumi',
  name: redisSubnetGroupName,
  tags: additionalDefaultTags,
  subnetIds: eksVpc.privateSubnetIds,
});

export const redisSecurityGroup = new aws.ec2.SecurityGroup('redisSecurityGroup', {
  description: 'Redis Security Group',
  name: 'redisSecurityGroup',
  tags: additionalDefaultTags,
});

export const userRedis = new aws.elasticache.ReplicationGroup(redisName, {
  autoMinorVersionUpgrade: true,
  automaticFailoverEnabled: true,
  engineVersion: redisEngineVersion,
  maintenanceWindow: redisMaintenanceWindow,
  multiAzEnabled: false,
  networkType: 'ipv4',
  nodeType: redisNodeType,
  numCacheClusters: redisNumCacheClusters,
  parameterGroupName: 'default.redis7.cluster.on', // Use the parameter group that matches the engine version
  port: 6379,
  replicationGroupId: redisName,
  securityGroupIds: [redisSecurityGroup.id],
  snapshotWindow: '07:00-08:00',
  tags: additionalDefaultTags,
  description: 'Managed by Pulumi',
}, 
{
  dependsOn: [
    eksCluster, 
    cniPolicyAttachment,
    ec2ContainerpolicyAttachment,
    nextgenNodePolicyAttachment, 
    workerNodepolicyAttachment,
  ],
  protect: false,
});

But you don't need those resources to have fixed names. You are passing (e.g.) redisSecurityGroup.id into the ReplicationGroup constructor, so the security group doesn't need to have a name parameter. So don't configure it.

giving it a new try now thanks for the tip

Nice. Well if anything stumps you, ask a new question, I'm here for a few more hours, probably.

ty I'm just plugging along