Changelogs » Pulumi

Pulumi

0.20.3

Major changes

-   None

Improvements

-   None

Bug fixes

-   None

0.20.2

Major changes

-   None

Improvements

-   Allow awaiters to be skipped by setting an annotation (https://github.com/pulumi/pulumi-kubernetes/pull/417)
-   Set managed-by: pulumi label on all created resources (https://github.com/pulumi/pulumi-kubernetes/pull/418)
-   Clean up docstrings for Helm package (https://github.com/pulumi/pulumi-kubernetes/pull/396)
-   Support explicit `deleteBeforeReplace` (https://github.com/pulumi/pulumi/pull/2415)

Bug fixes

-   Fix an issue with variable casing (https://github.com/pulumi/pulumi-kubernetes/pull/412)
-   Use modified copy of memcache client (https://github.com/pulumi/pulumi-kubernetes/pull/414)

0.20.1

Bug fixes

-   Fix namespace handling regression (https://github.com/pulumi/pulumi-kubernetes/pull/403)
-   Nest Input<T> inside arrays (https://github.com/pulumi/pulumi-kubernetes/pull/395)

0.20.0

Major changes

-   Add support for first-class Python providers (https://github.com/pulumi/pulumi-kubernetes/pull/350)
-   Upgrade to client-go 0.10.0 (https://github.com/pulumi/pulumi-kubernetes/pull/348)

Improvements

-   Consider PVC events in Deployment await logic (https://github.com/pulumi/pulumi-kubernetes/pull/355)
-   Improve info message for Ingress with default path (https://github.com/pulumi/pulumi-kubernetes/pull/388)
-   Autogenerate Python casing table from OpenAPI spec (https://github.com/pulumi/pulumi-kubernetes/pull/387)

Bug fixes

-   Use `node-fetch` rather than `got` to support Node 6 (https://github.com/pulumi/pulumi-kubernetes/pull/390)
-   Prevent orphaned resources on cancellation during delete (https://github.com/pulumi/pulumi-kubernetes/pull/368)
-   Handle buggy case for headless Service with no port (https://github.com/pulumi/pulumi-kubernetes/pull/366)

0.19.0

Major changes

-   Implement incremental status updates for `StatefulSet`
(https://github.com/pulumi/pulumi-kubernetes/pull/307)
-   Allow the `pulumi/kubernetes` YAML API to understand arbitrary URLs
(https://github.com/pulumi/pulumi-kubernetes/pull/328)
-   Add support for `.get` on CustomResources
(https://github.com/pulumi/pulumi-kubernetes/pull/329)
-   Add support for `.get` for first-class providers
(https://github.com/pulumi/pulumi-kubernetes/pull/340)

Improvements

-   Fix Ingress await logic for ExternalName Services
(https://github.com/pulumi/pulumi-kubernetes/pull/320)
-   Fix replacement logic for Job
(https://github.com/pulumi/pulumi-kubernetes/pull/324 and https://github.com/pulumi/pulumi-kubernetes/pull/324)
-   Fix Cluster/RoleBinding replace semantics
(https://github.com/pulumi/pulumi-kubernetes/pull/337)
-   Improve typing for `apiVersion` and `kind`
(https://github.com/pulumi/pulumi-kubernetes/pull/341)

0.18.0

Major changes

-   Allow Helm Charts to have `pulumi.Input` in their `values`
(https://github.com/pulumi/pulumi-kubernetes/pull/241)

Improvements

-   Retry REST calls to Kubernetes if they fail, greatly improving resiliance against resorce
operation ordering problems.
-   Add support for creating CRDs and CRs in the same app
(https://github.com/pulumi/pulumi-kubernetes/pull/271,
https://github.com/pulumi/pulumi-kubernetes/pull/280)
-   Add incremental await for logic for `Ingress`
(https://github.com/pulumi/pulumi-kubernetes/pull/283)
-   Allow users to specify a Chart's source any way they can do it from the CLI
(https://github.com/pulumi/pulumi-kubernetes/pull/284)
-   "Fix" "bug" that cases Pulumi to crash if there is a duplicate key in a YAML template, to conform
with Helm's behavior (https://github.com/pulumi/pulumi-kubernetes/pull/289)
-   Emit better error when the API server is unreachable
(https://github.com/pulumi/pulumi-kubernetes/pull/291)
-   Add support for Kubernetes v0.12.\* (https://github.com/pulumi/pulumi-kubernetes/pull/293)
-   Fix bug that spuriously requires `.metadata.name` to be specified in Kubernetes list types
(_e.g._, `v1/List`) (https://github.com/pulumi/pulumi-kubernetes/pull/294,
https://github.com/pulumi/pulumi-kubernetes/pull/296)
-   Add Kubernetes v0.13.\* support (https://github.com/pulumi/pulumi-kubernetes/pull/306)
-   Improve error message when `Service` fails to initialized
(https://github.com/pulumi/pulumi-kubernetes/pull/309)
-   Fix bug that causes us to erroneously report `Pod`'s owner
(https://github.com/pulumi/pulumi-kubernetes/pull/311)

0.16.15

Improvements

- When trying to `stack rm` a stack managed by pulumi.com that has resources, the error message now informs you to pass `--force` if you really want to remove a stack that still has resources under management, as this would orphan these resources (fixes [pulumi/pulumi2431](https://github.com/pulumi/pulumi/issues/2431)).
- Enabled Python programs to delete resources in parallel (fixes [pulumi/pulumi2382](https://github.com/pulumi/pulumi/issues/2382)). If you are using Python 2, you should upgrade to Python 3 or else you may experience problems when deleting resources.
- Fixed an issue where Python programs would occasionally fail during preview with errors about empty IDs being passed
to resources. ([pulumi/pulumi2450](https://github.com/pulumi/pulumi/issues/2450))
- Return an error from `pulumi stack tag` commands when using the `--local` mode.

0.16.14

Improvements

- Fix a regression in `pulumi/pulumi` introduced by 0.16.13 where an update could fail with an error like:


Diagnostics:
pulumi:pulumi:Stack (my-great-stack):
TypeError: resproto.InvokeRequest is not a constructor
at Object.<anonymous> (.../node_modules/pulumi/pulumi/runtime/invoke.js:58:25)
at Generator.next (<anonymous>)
at fulfilled (.../node_modules/pulumi/pulumi/runtime/invoke.js:17:58)
at <anonymous>


We appologize for the regression.  (fixes [pulumi/pulumi2414](https://github.com/pulumi/pulumi/issues/2414))

Improvements

- Individual resources may now be explicitly marked as requiring delete-before-replace behavior. This can be used e.g. to handle explicitly-named resources that may not be able to be replaced in the usual manner.

0.16.13

Major Changes

- When used in conjuction with the latest versions of the various language SDKs, the Pulumi CLI is now more precise about the dependent resources that must be deleted when a given resource must be deleted before it can be replaced (fixes [pulumi/pulumi2167](https://github.com/pulumi/pulumi/issues/2167)).

**NOTE**: As part of the above change, once a stack is updated with v0.16.13, previous versions of `pulumi` will be unable to manage it.

Improvements

- Issue a more perscriptive error when using StackReference and the name of the stack to reference is not of the form `<organization>/<project>/<stack>`.

0.16.12

Major Changes

- When using the cloud backend, stack names now must only be unique within a project, instead of across your entire account. Starting with version of 0.16.12 the CLI, you can create stacks with duplicate names. If an account has multiple stacks with the same name across different projects, you must use 0.16.12 or later of the CLI to manage them.

**BREAKING CHANGE NOTICE**: As part of the above change, when using the 0.16.12 CLI (or a later version) the names passed to `StackReference` must be updated to be of the form (`<organization>/<project>/<stack>`) e.g. `acmecorp/infra/dev` to refer to the `dev` stack of the `infra` project in the `acmecorp` organization.

Improvements

- Add `--json` to `pulumi config`, `pulumi config get`, `pulumi history` and `pulumi plugin ls` to request the output be in JSON.

- Changes to `pulumi new`'s output to improve the experience.

0.16.11

Improvements

- In the nodejs SDK, `pulumi.interpolate` and `pulumi.concat` have been added as convenient ways to combine Output values into strings.

- Added `pulumi history` to show information about the history of updates to a stack.

- When creating a project with `pulumi new` the generated `Pulumi.yaml` file no longer contains the template section, which was unused after creating a project

- In the Python SDK, the `is_dry_run` function just always returned `true`, even when an update (and not a preview) was being preformed. This has been fixed.

- Python programs will no longer deadlock due to exceptions in functions run during applies.

0.16.10


      

0.16.9


      

0.16.8

Improvements

- Updated to the v1.58.0 version of the AWS Terraform Provider.

- Examples in the documentation comments for the Node.js SDK have improved formatting.

- Support for the `deleteBeforeReplace` resource option and improved
delete-before-replace behaviour introduced in [Pulumi
0.16.14](https://github.com/pulumi/pulumi/blob/master/CHANGELOG.md01614-released-january-31st-2019).

0.16.7

Improvements

- Updated to the v1.56.0 version of the AWS Terraform Provider.

- Documentation comments for the Node.js SDK now include examples

0.16.6

Improvements

- Auto name `aws.elasticache.ReplicationGroup.replicationGroupId` (fixes [pulumi/pulumi-aws426](https://github.com/pulumi/pulumi-aws/issues/426)).

- Fix a deprecation warning when using `aws.wafregional.ByteMatchSet` (fixes [pulumi/pulumi-aws429](https://github.com/pulumi/pulumi-aws/issues/429)).

- Updated package constraints such that we do not depend on unreleased versions of `pulumi/pulumi`.

- Add some more constants for EC2 instance types.  Thanks to [BradBarnich](https://github.com/BradBarnich)!

0.16.5

- Supply easy mechanisms to add Internet and NAT gateways to a VPC.
- Change awsx.elasticloadbalancingv2.Listener.endpoint from a method to a property.
- Change awsx.apigateway.ProxyRoute.target to be a richer type to allow extensibility.
- Allow awsx.elasticloadbalancingv2.NetworkListener to be used as ProxyRoute.target to simply
incoming APIGateway routes to a NetworkListener endpoint.
- Add support for arbitrary APIGateway integration routes (i.e. to any supported aws service).
Note: this comes with a small breaking change where the names of some apigateway types have
changed from ProxyXXX to IntegrationXXX.

0.16.4

- Renamed 'aws-infra' package to 'awsx'.
- Moved `aws.apigateway.x.Api` from `pulumi/aws` into this package under the name `awsx.apigateway.Api`.

0.16.3

- Experimental abstractions have been promoted to supported abstractions.  see new modules for:
- autoscaling
- ec2
- ecs
- elasticloadbalancingv2

0.16.2

Improvements

- Add some experimental abstractions for Services and Tasks in the `experimental` module.

0.16.1

Improvements

- Fix an issue where passing a cluster to another component would fail in some cases.