CHANGELOG.md

Unreleased

v5.35.0

• Add heartbeat monitoring options to alert source resource (interval, grace_period_seconds, failure_threshold)

v5.34.0

• Add support for delay nodes on escalation paths

v5.33.0

• Add alert_events_url to alert source resource and data source

v5.32.0

• Add incident_maintenance_window resources to create maintenance windows

v5.31.0

• Add support for repeat_config on escalation paths

v5.30.0

• Add support for emoji on alert attributes

v5.29.0

• Add auto_resolve_incident_alerts and auto_resolve_timeout_minutes to alert source provider

v5.28.0

• Added support for merge_strategy on alert attributes

v5.27.0

• Add support for message_template on alert routes

v5.26.2

• Add plan-time validation to the branches operations in expressions

v5.26.1

• Fix error when workflow steps are upgraded

v5.26.0

• Fixes an issue where you couldn't destroy catalog entries where the type was managed and you were only managing a subset of its attributes
• Fixes an issue where incident_template.severity was being required, when it's actually optional

v5.25.0

• Add support for owning_team_ids for alert sources and alert routes

v5.24.1

• Fix schema_only drift when catalog type attribute mode changes

v5.24.0

• Add support for private alert sources

v5.23.1

• Fix handling of empty external_id in catalog entries

v5.23.0

• Add "include_private_escalations" option to workflows

v5.22.0

• Add "incident_escalation_path" data source for getting escalation paths by id or name

v5.21.1

• Revert expressions to use a set type. There order isn't consistent when coming back from the server.

v5.21.0

• Improve terraform plan performance by using a list type rather than set for conditions, condition_groups and expressions. This may cause a one-time ordering changes in plans, this is expected and will resolve after applying.

v5.20.0

• Improve terraform plan performance by using a list type rather than set for array_values. This may cause a one-time ordering changes in plans, this is expected and will resolve after applying.
• Add support for auto_relate_grouped_alerts for Alert Routes

v5.19.1

• Performance improvement: incident_catalog_entries now uses bulk update API to batch updates in groups of 100, significantly reducing API calls for large catalog syncs

v5.19.0

• Automatically remove resources from Terraform state if they're not found remotely (contribution from @maxtacu)

v5.18.0

• We will now allow you to use dynamic variables for working_intervals in incident_schedule.

v5.17.1

• Handle 404s when an alert attribute has been deleted outside of Terraform (contribution from @maxtacu)

v5.17.0

• Fix rate limiting by adding missing unit to backoff

v5.16.0

• Add error logging for panics

v5.15.0

• Add debug logging for outbound HTTP requests
• Fixed a bug where catalog type attributes weren't defaulting to being marked as managed via Terraform, unless explicitly marked. You may see a change in the state of schema_only in your next plan if you hadn't set it previously.

v5.14.0

• Fix issue with HTML-like characters (e.g. >) in JSON engine values being incorrectly encoded
• Clarify that source_repo_url is required for catalog types in documentation
• Update dependencies to latest versions

v5.13.0

• Support for custom HTTP alert source configuration
• Allow configuration of ack_mode for escalation paths
• Improve description for use_name_as_identifier field in catalog types to be clearer
• Improve escalation path resource description to be more focused

v5.12.0

• Fix incident_catalog_type to be able to better handle undefined or empty category lists
• Expose incident_catalog_entries in the documentation
• No longer escape HTML characters in engine JSON strings
• incident_alert_attribute now has an optional required property. Set this to true for attributes which you expect to be present on all alerts.

v5.11.0

• Add incident_catalog_entries data source to get catalog entries for a specific catalog type. This is useful for building up a list of catalog entries which you might be managing via catalog-importer.

v5.10.0

• Enforce consistent ordering of keys in engine literal values which are JSON objects.
• Made handovers on incident_schedule.config.rotations.* required, as it was always required by the API, but was not marked as such in the provider.
• Switch runs_on_incident_modes on workflow resources to be a set and not a list, so it's not sensitive to ordering.
• Add incident_schedule data source for retrieving existing schedules by ID or name.

v5.9.1

• Improve documentation to reflect potential values of enumerated values.
• Improve example for incident_alert_route documentation.

v5.9.0

• Update the documentation for incident_alert_source, incident_alert_route, incident_escalation_path and incident_schedule to reference the 'Export' flow in the dashboard
• Adds incident_alert_sources as a plural data source, to retrieve a list of your alert sources

v5.8.0

• incident_schedule_resource now uses sets for rotations as the ordering of them does not matter.
• Alert sources and alert routes created by Terraform or imported to Terraform will be tagged as such and won't be editable in the incident.io dashboard.

v5.7.1

• Support up to 3 levels of branch nesting on escalation paths

v5.7.0

• Improve the documentation for team_ids in incident_escalation_path
• incident_alert_source supports dynamic values for all attributes - for example initialising attributes from local variables.

v5.6.0

• Improve the documentation for channel_config in incident_alert_route
• Fix a bug where empty slices of team_ids would be sent to the API as null
• Fix a bug where empty slices of managed_attributes would mean we mark every attribute as managed in Terraform, whereas it should mean every attribute is managed in the dashboard.

v5.5.0

• Add grouping_window_seconds to alert route incident config. This is a required field that was being defaulted to 0, meaning any alert route created through terraform ended up with grouping disabled.
• Fix a bug where custom fields would show a diff when specified in a different order to when the custom field itself was created. As ordering does not matter, this now uses a set rather than a list.
• Make incident_template required on alert routes. This was previously marked optional, but our provider would crash if it was not supplied. This is also required by our API, so we have made it required in the provider.

Note that we've decided to release this as a minor version despite the breaking change of grouping_window_seconds being required. This is because the field was previously defaulted to 0, and so any alert route created through terraform would have had grouping disabled. As such, we consider this a bug fix and encourage all users to upgrade.

If you want to leave grouping disabled, set grouping_window_seconds to 0.

v5.4.2

• Add validation for RFC3339 timestamp format in handover_start_at and effective_from fields to prevent invalid dates
• incident_alert_route supports dynamic values for all attributes - for example enabling channel_config using a variable.
• Fixed a bug where the plan for incident_alert_route would always show a diff for incident_template.name.array_value and incident_template.summary.array_value.

v5.4.1

• Allow terraform import for incident_catalog_type_attribute
• Add terraform import support to the documentation

v5.4.0

• Adds incident_alert_route resource for managing alert routes.

v5.3.2

• In the catalog entry resource, we now guard against cases where the type of attribute_values is inferred to be unknown during the validation of managed attributes.

v5.3.1

When loading workflows, ensure that any additional parameter bindings are skipped, so that apply does not see these as differences.

v5.3.0

The incident_schedule and incident_escalation_path resources now support a team_ids attribute to associate those resources with a team.

v5.3.0-beta1

The incident_schedule and incident_escalation_path resources now support a team_ids attribute to associate those resources with a team.

v5.2.0

Temporarily remove team_ids support.

v5.1.0

The incident_schedule and incident_escalation_path resources now support a team_ids attribute to associate those resources with a team.

v5.0.0

Breaking changes

incident_catalog_type's source_repo_url attribute is now required.

This prevents the catalog type from being edited manually, and ensures there is a link from the incident.io dashboard to the configuration that defines the catalog type.

Schema-only attributes

Sometimes you want to define most of a catalog entry's attributes in Terraform, but allow other attributes to be edited in the incident.io dashboard.

This is now possible with schema-only attributes:

• add schema_only = true to the incident_catalog_type_attribute resource to mark the attribute as schema-only: it will be created by Terraform, but values can be edited in the incident.io dashboard.

• add managed_attributes to the incident_catalog_entry or incident_catalog_entries, and specify only the attributes that should be managed in Terraform.

  By excluding schema-only attributes from this list, changes to those attribute values made in the dashboard will not cause unnecessary diffs when you next run terraform plan.

New data sources

There are now data sources for incident_catalog_type_attribute and incident_catalog_entry.

These allow you to look up an attribute by its catalog type ID and name, and an entry by its catalog type ID and name, external ID, or alias.

This is useful for:

• managing entries of a catalog type across multiple modules: you can use the incident_catalog_type_attribute data source to get the ID of attributes, without needing to pass the ID between modules.
• using data from your catalog elsewhere in Terraform: for example attributes of your Team catalog type.

v4.3.3

• Fixed a panic when using an incident_escalation_path with a notify_channel node that has a time_to_ack_interval_condition set.

v4.3.2

• The order of attributes set within an alert source's template block is now ignored when planning and applying changes.

v4.3.1

• Mark the 'email_address' and 'secret_token' fields on incident_alert_source as remaining the same for the lifetime of an alert source, to avoid misleading plans.

v4.3.0

• incident_custom_field resource now supports catalog-powered custom fields, including controlling which attribute is used to group options, add helptext, and filter the available options by another field's value.
• incident_custom_field data source exposes extra attributes for catalog-powered custom fields.

4.2.0

• incident_alert_attribute resource and data source, for managing your alert attributes
• incident_alert_source resource for managing your alert sources
• A number of dependencies have been updated.

4.1.0

• Escalation paths created by Terraform or imported to Terraform will be tagged as such and won't be editable in the incident.io dashboard.

4.0.4

• Updates the documentation for custom fields

4.0.3

• Updates to documentation

4.0.2

• Adds support for adding slack_channel nodes to escalation paths

4.0.1

• Ensures that client operations will fail with errors when an endpoint would otherwise have returned 204 no content for a successful operation.

4.0.0

• Fixes an issue where the provider might fail to import Terraform state for a schedule with working hours applied

To upgrade to v4.0.0, if you've got on-call schedules with working hours specified in your Terraform code, you'll need to rename the following properties of your working_intervals:

• day -> weekday
• start -> start_time
• end -> end_time

3.8.11

• Add external_id to resource_catalog_entry

3.8.10

• Fix another issue with condition group arrays that was producing inconsistent apply results

3.8.9

• Fix a bug with the serialisation of empty condition group arrays that caused validation errors

3.8.8

• Support workflow shortforms for triggering manual workflows.
• Fix regression from 3.8.6 that impacted creating and updating schedules with working intervals

3.8.7

• Migrate to a new internal client, no functional changes.

3.8.6

• Add support for the incident_incident_role data source.

3.8.5

• Fixed an issue (#99) where the provider crashed if a round robin config with no minutes was provided

3.8.4

• Add support for holidays_public_config on the incident_schedule resource

3.8.3

• Retry on 429 responses from the API, respecting the Retry-After header

3.8.2

• Fixed incorrect produced an unexpected new value errors when configuring escalation paths

3.8.1

• Improved handling of HTTP errors

3.8.0

• Add support for schedule_mode on the incident_escalation_path resource target parameter
• Add support for round_robin_config on the incident_escalation_path resource level parameter

3.7.0

• Add support for path attributes on the incident_catalog_type_attribute resource
• Add support for categories on the incident_catalog_type resource

3.6.0

• incident_escalation_path for configuring escalation paths.

3.5.0

• data sources for incident_custom_field and incident_custom_field_option, contributed by @mdb

3.4.0

• data source for incident_catalog_type to allow for lookups of catalog types

3.3.1

• Docs update to include examples of incident_workflow resource

3.3.0

• Add support for workflows using the incident_workflow resource

3.2.3

• Docs update to include examples of incident_schedule resource

3.2.2

• Adds supports for on-call schedules using the incident_schedule resource
• Adds support for user looksups using the incident_user data source

3.2.1

• Add support for setting the source_repo_url on catalog types
• Fix a bug where we'd panic if we received a specific kind of error when updating catalog entries

3.2.0

• Add support for backlink attributes on catalog types

3.1.2

• Marks type_name as requiring a replace, as it is immutable
• Updates our docs so they are a lot clearer on how to connect attributes

3.1.1

• Handle 404 for all resources without panicking, and remove resource from state

3.1.0

• Add support for setting the type_name of a catalog type. This allows other catalog attributes to refer to this type by a friendly name, rather than the randomly generated ID

3.0.0

• Remove SemanticType from catalog types (This has never been used by our application, so we've decided to remove it from the provider as we have no plans to use it.)
• Move to CustomFieldsV2 API as we are deprecating a number of fields from the CustomFieldsV1 API (required, show_before_closure, show_before_creation, show_before_update, show_in_announcement_post). These will now be controlled via 'Incident Forms' which (for now) will only be available via the web dashboard. This will enable users to have much more control over the way they configure their incident forms.
• Move to IncidentRolesV2 API as we are deprecating the required field from the IncidentRolesV1 API. This will now be controlled via 'Incident Forms' which (for now) will only be available via the web dashboard. This will enable users to have much more control over the way they configure their incident forms.

To upgrade to v3, you will need to remove the deprecated fields from any custom_field and incident_role resources. You'll also need to remove any references to semantic_type

2.0.2

• Handle omission of empty list or null array_value in catalog entries (#36)

2.0.1

• Update client to latest API schema
• Remove any disclaimers about the catalog being in beta ahead of launch

2.0.0

• Rename alias in catalog_entries and catalog_entry to aliases in support for multiple alias entries
• Handle catalog types having been removed without panicking

1.4.3

• Handle 404 for catalog types without panicking

1.4.2

• Fix bug in framework patch that meant we never defaulted our log level

1.4.1

• Pin the correct dependency to include our logging patch

1.4.0

• incident_catalog_entries for large entry counts

1.3.1

• Fix bug around omitted empty arrays

1.3.0

• Support alias and rank for catalog_entry

1.2.0

• Technically new feature, this represents attribute values on catalog entries as sets to avoid unnecessary diffs when reordering the attributes

1.1.0

• Adds support for catalog types, attributes and entries

1.0.3

• Bugfix for terraform provider variables

1.0.2

• Fix API key setting via provider attribute
• Provide user-agent of terraform-provider-incident/version for all requests
• Fix creating severities without providing a rank

1.0.1

• Severity rank is computed (#2)

1.0.0

Initial release, including support for:

• Custom fields (incident_custom_field)
• Custom field options (incident_custom_field_option)
• Incident roles (incident_incident_role)
• Severities (incident_severity)
• Statuses (incident_status)