MongoDB Atlas Terraform Provider

The following integration patterns are available in the examples directory:

The mongodbatlas_advanced_cluster resource allows you to create, edit, and delete advanced MongoDB Atlas clusters.

Best Practices:

• It is highly recommended to use the Official MongoDB Atlas Cluster Module instead of managing the raw resource, as the module implements best practices by default.
• If migrating from provider version v1.x.x to v2.0.0 or later, you must follow the migration guide to update your configuration.

Key Capabilities:

• Supports creating Flex clusters.
• Supports upgrading M0 clusters to Flex.
• Supports upgrading Flex clusters to Dedicated clusters.
• Creates a network container for every provider/region combination specified. You can access these via the computed replication_specs[#].container_id attribute.

The mongodbatlas_event_trigger resource allows you to automate actions in MongoDB Atlas App Services by responding to specific events. You can create three types of triggers:

1. DATABASE: Responds to changes in MongoDB collections (e.g., INSERT, UPDATE).
2. AUTHENTICATION: Responds to user authentication events (e.g., LOGIN, CREATE, DELETE).
3. SCHEDULED: Runs on a specific schedule defined by a cron expression.

Important Note: Changing the app_id will force a replacement of the resource, deleting the trigger from the old Atlas App Services app and creating it in the new one.

Note on Functions: While you can create database triggers via Terraform, you cannot create the Atlas Functions themselves using this provider. You must define the functions in Atlas separately.

resource "mongodbatlas_event_trigger" "test" {
  project_id = "PROJECT ID"
  app_id = "APPLICATION ID"
  name = "NAME OF THE TRIGGER"
  type = "DATABASE"
  # ... other configuration ...
}

When upgrading mongodbatlas_advanced_cluster from v1.x to v2.0.0, several breaking changes occur in the configuration schema:

Removed Attributes

• id (Use cluster_id instead)
• disk_size_gb
• replication_specs.#.num_shards
• replication_specs.#.id
• advanced_configuration.default_read_concern
• advanced_configuration.fail_index_key_too_long

Structural Changes

• Sharding: Supports only the new sharding configuration for independent shard scaling.
• List Attributes: replication_specs and region_configs have changed from blocks to list attributes (arrays of objects).
• Single Attributes: connection_strings, timeouts, advanced_configuration, bi_connector_config, pinned_fcv, electable_specs, read_only_specs, analytics_specs, auto_scaling, and analytics_auto_scaling are now single attributes (objects) instead of blocks. Remove indices like [0] or .0 when referencing them.
• Maps: tags and labels are now maps instead of blocks.

# Old tags block syntax
tags {
  key   = "env"
  value = "dev"
}

# New tags map syntax
tags = {
  env           = "dev"         # key strings without blanks can be enclosed in quotes but not required
  "tag 2"       = "val"         # enclose key strings with blanks in quotes
  (var.tag_key) = "another_val" # enclose key expressions in brackets to evaluate
}

The recommended way to manage auto-scaling without encountering Terraform plan drift is to set use_effective_fields = true in the mongodbatlas_advanced_cluster resource.

This approach allows you to define an initial instance_size in your configuration, while Terraform uses the 'effective' values (the actual sizes Atlas has scaled to) to prevent unnecessary updates. You can use the mongodbatlas_advanced_cluster data source with use_effective_fields = true to read these actual values back into your configuration.

Module Authors: This is highly recommended for reusable modules to support both auto-scaling and non-auto-scaling clusters without requiring lifecycle blocks.

resource "mongodbatlas_advanced_cluster" "test" {
  project_id            = var.project_id
  name                  = "AutoScalingCluster"
  cluster_type          = "SHARDED"
  use_effective_fields  = true

  replication_specs = [{
    region_configs = [{
      electable_specs = {
        instance_size = "M40"
        node_count    = 3
      }
      auto_scaling = {
        compute_enabled = true
        compute_max_instance_size = "M60"
      }
      # ...
    }]
    zone_name = "Zone 1"
  }]
}

# Read effective values to see what Atlas has scaled to
data "mongodbatlas_advanced_cluster" "test" {
  project_id           = mongodbatlas_advanced_cluster.test.project_id
  name                 = mongodbatlas_advanced_cluster.test.name
  use_effective_fields = true
  depends_on           = [mongodbatlas_advanced_cluster.test]
}

output "shard_sizes" {
  value = [
    for idx, spec in data.mongodbatlas_advanced_cluster.test.replication_specs : {
      shard_index           = idx
      configured_size       = spec.region_configs[0].electable_specs.instance_size
      actual_electable_size = spec.region_configs[0].effective_electable_specs.instance_size
      actual_analytics_size = spec.region_configs[0].effective_analytics_specs.instance_size
    }
  ]
}