Skip to content

Latest commit

 

History

History
210 lines (165 loc) · 6.47 KB

MIGRATION.md

File metadata and controls

210 lines (165 loc) · 6.47 KB
Raw

Migration Guide to v1.0.0

For everyone upgrading from a version before v1.0.0, we recommend upgrading straight to v1.2 or higher.

Important

If you have created checks in 1.0.x or 1.1.x you will first need to upgrade to 1.1.4 before jumping to >= 1.2.5

We at OpsLevel have been working on upgrading our Terraform provider to version 1.0.0.

While the majority of these improvements are under the hood, there are a few Terraform configuration changes to be aware of.

For errors encountered during terraform apply, see the State Migration section below - Applies only to v1.0.0 up to v1.1.x

BREAKING CHANGES to some resource fields

The following resources with fields that were “block type fields" will need an “=” added.

  • opslevel_check_manual
# OLD - no "=" sign
resource "opslevel_check_manual" "example" {
  update_frequency { ... }
}
# New has "=" sign
resource "opslevel_check_manual" "example" {
  update_frequency = { ... }
}

Just like opslevel_check_manual.update_frequency, the following opslevel_resource.fields will need an "=" added.

  • opslevel_check_alert_source_usage.alert_name_predicate
  • opslevel_check_repository_file.file_contents_predicate
  • opslevel_check_repository_grep.file_contents_predicate
  • opslevel_check_repository_search.file_contents_predicate
  • opslevel_check_service_ownership.tag_predicate
  • opslevel_check_service_property.predicate
  • opslevel_infrastructure.provider_data
  • data.opslevel_repositories.filter (datasource)
  • data.opslevel_services.filter (datasource)

BREAKING CHANGES with OpsLevel's updated "plural" datasources

Many of our datasources were updated to provide a better user experience and may require a few small updates.

Terraform configuration changes during upgrade

When upgrading from versions older than 0.12.0-0, updating the outputs of "plural" datasources may be needed.

# Requesting the datasource remains the same
data "opslevel_domains" "all" {}

# OLD - before v0.12.0-0
output "all" {
  value = data.opslevel_domains.all
}
output "domain_names" {
  value = data.opslevel_domains.all.names
}

# NEW - versions v0.12.0-0 and after
output "all" {
  value = data.opslevel_domains.all.domains
}
output "domain_names" {
  value = data.opslevel_domains.all.domains[*].name
}

This same pattern applies to data.opslevel_scorecards, data.opslevel_services, data.opslevel_teams, etc.

Output Example - data.opslevel_domains with version >= 0.12.0

Given this configuration:

data "opslevel_domains" "all" {}

output "all" {
  value = data.opslevel_domains.all.domains
}

output "domain_names" {
  value = sort(data.opslevel_domains.all.domains[*].name)
}

Will return:

data.opslevel_domains.all: Reading...
data.opslevel_domains.all: Read complete after 1s

Changes to Outputs:
  + all          = [
      + {
          + aliases     = [
              + "alias_1",
              + "alias_2",
            ]
          + description = null
          + id          = "Z2lkOi8vb3BzbGV2ZWwvRW50aXR5T2JqZWN0LzEwOTk2NjU"
          + name        = "Big Domain"
          + owner       = "Z2lkOi8vb3BzbGV2ZWwvVGVhbS8xNDQwNg"
        },
      + {
          + aliases     = [
              + "example_1",
              + "example_2",
            ]
          + description = "Example description"
          + id          = "Z2lkOi8vb3BzbGV2ZWwvRW50aXR5T2JqZWN0LzE4ODg2NzA"
          + name        = "Example Domain"
          + owner       = "Z2lkOi8vb3BzbGV2ZWwvVGVhbS85NzU5"
        },
    ]
  + domain_names = [
      + "Big Domain",
      + "Example Domain",
    ]

Output Example - data.opslevel_domains before version 0.12.0-0

Given this configuration:

data "opslevel_domains" "all" {}

output "all" {
  value = data.opslevel_domains.all
}

output "domain_names" {
  value = sort(data.opslevel_domains.all.names)
}

Will return:

data.opslevel_domains.all: Reading...
data.opslevel_domains.all: Read complete after 1s

Changes to Outputs:
  + all          = [
      + aliases      = [
          + "alias_1",
          + "alias_2",
          + "example_1",
          + "example_2",
        ]
      + descriptions = [
          + "",
          + "Example description",
        ]
      + ids          = [
          + "Z2lkOi8vb3BzbGV2ZWwvRW50aXR5T2JqZWN0LzEwOTk2NjU",
          + "Z2lkOi8vb3BzbGV2ZWwvRW50aXR5T2JqZWN0LzE4ODg2NzA",
        ]
      + names        = [
          + "Big Domain",
          + "Example Domain",
        ]
      + owners       = [
          + "Z2lkOi8vb3BzbGV2ZWwvVGVhbS8xNDQwNg",
          + "Z2lkOi8vb3BzbGV2ZWwvVGVhbS85NzU5",
        ]
    }
  + domain_names = [
      + "Big Domain",
      + "Example Domain",
    ]

State Migration (Applies to v1.0.0 up to v1.1.x)

We recommend upgrading straight to v1.2 or higher. But for versions between v1.0.0 and v1.1.x, the following may be needed.

The structure of some resource data has changed from v0.11.0 to v1.0.0. If you encounter erros like Error: Unable to Read Previously Saved State for UpgradeResourceState while upgrading to the OpsLevel v1.0.0 Terraform provider, the following will get you unstuck.

Get ids of resources that raise errors

Assuming opslevel_check_tag_defined.my_tag raises an error like AttributeName("environment_predicate"): invalid JSON, expected "{", got "[", run terraform state show opslevel_check_tag_defined.my_tag and extract the value of id. Extract and keep all ids for resources that raise errors.

If terraform state show ... raises an error, try terraform state pull > my_tf_state.json and extract ids from my_tf_state.json.

Remove problematic resources from state

This does NOT remove the resources from OpsLevel - only from Terraform's state. Run terraform state rm opslevel_check_tag_defined.my_tag and repeat for all resources that raise errors. Optionally verify resources removed from state with terraform state list.

Import resources from OpsLevel into Terraform state

Import OpsLevel resources into Terraform's state by running terraform import opslevel_check_tag_defined.my_tag <resource-id> and repeat for all resources that raise errors. Note matches the id of the OpsLevel resource as returned by the OpsLevel cli, if id is not otherwise known. Optionally verify resources imported to state with terraform state list.