* fix: filter empty metrics before we collect last seen toggles.
fixes: #2104
* fix: add a last-seen service to batch last-seen toggle updates
Co-authored-by: Fredrik Strand Oseberg <fredrik.no@gmail.com>
* fix: PATs should have an unique description
* add pat validation on the back-end service
* Update src/lib/services/pat-service.ts
Co-authored-by: Simon Hornby <liquidwicked64@gmail.com>
* fix: only consider current user's PATs
* fix tests
* cleanup
* Update frontend/src/component/user/Profile/PersonalAPITokensTab/CreatePersonalAPIToken/CreatePersonalAPIToken.tsx
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
* Update src/test/e2e/api/admin/user/pat.e2e.test.ts
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
Co-authored-by: Simon Hornby <liquidwicked64@gmail.com>
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
* fixed segments not being copied
* fix fmt
* bug fix
* return segmentId[] when getting a feature strategy
* do not return segments if they are not there
* Update src/lib/services/feature-toggle-service.ts
Co-authored-by: Fredrik Strand Oseberg <fredrik.no@gmail.com>
* fix lint
* fix: more explicit column sorting and bug fix
* update snapshot
* rollback
* add segment ids to feature strategies
* bug fix
Co-authored-by: Fredrik Strand Oseberg <fredrik.no@gmail.com>
## What
This PR does two things:
1. Add tests to ensure that all OpenAPI tags are listed in the root-level `tags` property of the spec.
2. Changes the tags for endpoints that use any of the deprecated tags (specifically `admin`).
## Why
When we moved to using the docusaurus OpenAPI doc generator, we made some changes to existing tags. Moving away from having just 'admin', 'client', and one or two more, we now use a more granular system, which makes it easier to find an endpoint by category.
However, there's currently nothing enforcing this and this bit of knowledge hasn't been clearly communicated to anyone. Because of this, we have some endpoints that are currently just grouped under a general 'admin' tag.
This has two drawbacks:
- In Swagger UI, all these endpoints are grouped together. This means that they break with the expectation that each endpoint has its own category. Further, being lumped together means that the 'admin' group is hard to read.
- The 'admin' tag is (on purpose) not included in the root-level `tags` property of the generated OpenAPI spec. This means that the OpenAPI docusaurus generator doesn't pick up those endpoints, which means that they're not included in the docs on docs.getunleash.io.
## How
By implementing tests that:
1. Check that the tags included in the spec are a subset of the "approved" tags
2. By checking each path in the OpenAPI spec and making sure that all its tags are listed in the root-level tag list. If this is not the case, the test fails and it logs an error explaining what is wrong to the console. An example of such a message is:
```
The OpenAPI spec contains path-level tags that are not listed in the root-level tags object. The relevant paths, operation ids, and tags are as follows:
POST /api/admin/feedback (operation id: createFeedback) has the following invalid tags: "admin"
PUT /api/admin/feedback/{id} (operation id: updateFeedback) has the following invalid tags: "admin"
For reference, the root-level tags are: "Addons", "Admin UI", "API tokens", "Archive", "Auth", "Client", "Context", "Edge", "Environments", "Events", "Features", "Import/Export", "Metrics", "Operational", "Playground", "Projects", "Public signup tokens", "Strategies", "Tags", "Unstable", "Users"
```
## Commits
* fix: ensure that all root-level tags in the spec are 'approved tags'
* fix: test that all listed path tags exist in the root-level tags
* fix: use "API tokens" tag for PAT endpoints
* fix: add comment explaining why tags test is there
* fix: Update snapshot
* fix: ensure that spec tags are a subset of the approved Tags
* fix: improve error message when tags don't match
* fix: further tweaks in log format
* fix: slack back link url to point to the feature page
* update snapshots
* update feature link in case project info is available in event
* update missing snapshot
* feat: new user dropdown and profile page
* fix: add popup boxShadow to dark-theme
* fix: update routes snap
* refactor: move some tab specific logic into tabs component
* add useProfile hook example
* fix profile tab header (no name)
* fix: hide user popup when clicking profile link
* - add PATs to profile;
- add route logic to profile;
- refactor TimeAgoCell title;
- misc fixes and refactoring;
* add profile info to profile tab
* simplify req paths
* add PAT flag to the front-end
* fix: some UI adjustments
* change user popup buttons to links
* fix profile on front-end, add role description
* update delete PAT text
* address some PR comments
* address PR comments
* some more UI fixes and refactoring
* move password request to API hook
* refactor: user creation screen cleanup
* feat: deprecation notice for google sso
* fix: docs openid typo
* invite link bar
* invite link page
* fix prettier docs
* regenerated openapi
* hooks for invite page api
* update openapi
* feat: invite link update
* feat: add public signup token soft-delete
* public signup frontend feature flag
* fix: new user api issues
* feat: allow for creating new user from invite link
* Feat/invite user public controller (#2106)
* added PublicInviteController for public urls
* added PublicInviteController for public urls
* added PublicInviteController for public urls
* added PublicInviteController for public urls
* fix test
* fix test
* update openapi
* refactor: password reset props
* fix: public invite schema and validation
Co-authored-by: Fredrik Strand Oseberg <fredrik.no@gmail.com>
* user invite frontend
Co-authored-by: Fredrik Strand Oseberg <fredrik.no@gmail.com>
* invite link delete confirmation dialog
* refactor: password reset action
* fix: new user invite loading state
* fix: run ts check with ci
* revert openapi changes
* fix: invite token api interface
* fix: openapi schema index
* fix: update test snapshots
* update frontend snapshot
* fix: prettier ci
* fix: updates after review
Co-authored-by: andreas-unleash <104830839+andreas-unleash@users.noreply.github.com>
Co-authored-by: Fredrik Strand Oseberg <fredrik.no@gmail.com>
* Middleware first version
* Middleware tests
* Add tests
* Finish middleware tests
* Add type for request
* Add flagresolver
* Fix snapshot
* Update flags and tests
* Put it back as default
* Update snapshot
* Add middleware test
* Middleware first version
* Middleware tests
* Add tests
* Finish middleware tests
* Add type for request
* Add flagresolver
* Fix snapshot
* Update flags and tests
* Put it back as default
* Update snapshot
* bug fix
* bug fix
* remove doc file
* store fixes
* bug fix
* rollback deleted file
* fix test
* add url to token
* return all tokens not just active
* add url fix
* PR comment
* PR comment
* PR comment
* add the flag to the experimental options
* fix env var name
## What
This change reverts the spec change to create-api-token-schema that
was introduced in 97c2b3c089
That commit changed the endpoint from allowing any string to only
allowing strings that match the values of the ApiTokenType value.
## Why
Because we already validate this on the backend, I didn't think it
would break anything. However, because OpenAPI enums are case
sensitive (and do not currently support anything else), it _did_ break
the API. ([stack overflow post that explains how this
works](https://stackoverflow.com/questions/60772786/case-insensitive-string-parameter-in-schema-of-openapi))
## Of note
This should have been caught by the tests, but it didn't show up until
it was merged to the main branch. I think it might be worth looking
into how we do the PR testing.
## What
This PR adds an extra layer of OpenAPI validation testing to what we already have. It also fixes any issues that make the new tests fail.
## Why
While the current OpenAPI validation takes care of _some_ things, there's also things it misses, as shown by #2055. By adding the OpenAPI Enforcer package, we should hopefully be able to catch more of these errors in the future. The enforcer _does_ flag the issue in #2055 as an error.
## How
By adding the OpenAPI Enforcer package and making whatever changes it picks up on.
By adding location headers to all our 201 endpoints. I also had to change some signatures on `create` store methods so that they actually return something (a lot of them just returned `void`).
## Discussion points
### Code changes
Some of the code changes may not be necessary or we may want to change more code to align with what changes have been done. It may be worth standardizing on a pattern for `*store.create` methods, so that they always return an identifier or the stored object, for instance.
### On relative URIs
The 201 location headers use relative URIs to point to the created resources. This seemed to be the easiest way to me, as we don't need to worry about figuring out what the absolute URL of the instance is (though we could probably just concat it to the request URI?). The algorithm for determining relative URIs is described in [RFC 3986 section 5](https://www.rfc-editor.org/rfc/rfc3986#section-5).
There's also some places where I'm not sure we _can_ provide accurate location url. I think they're supposed to point _directly at_ whatever the resource is, but for some resources (such as api tokens), you can only get it as part of a collection. From [RFC 9110 on the location field](https://httpwg.org/specs/rfc9110.html#field.location) (emphasis mine):
> the Location header field in a [201 (Created)](https://httpwg.org/specs/rfc9110.html#status.201) response is supposed to provide a URI that is **specific** to the created resource.
A link to a collection is not specific. I'm not sure what best to do about this.
### Inline comments
I've added a number of inline PR comments that I'd love to get some feedback on too. Have a look and let me know what you think!
### Unfinished business
I've added some juicy comments to some of the files here. They contain non-blocking issues that I'm tracking (via github issues). We should resolve them in the future, but right now they can stay as they are.
## Commits
* Feat: add openapi-enforcer + tests; fix _some_ issues
* Test: allow non-standard string formats
* validation: fix _some_ 201 created location header endpoints
* #1391: fix remaining 201 location headers missing
* Refactor: use the ajv options object instead of add* methods
* #1391: flag validation errors and warnings as test failures
* #1391: modify patch schema to specify either object or array
We don't provide many patch endpoints, so we _could_ create separate
patch operation objects for each one. I think that makes sense to do
as part of the larger cleanup. For now, I think it's worth to simply
turn it into one of these. While it's not entirely accurate, it's
better than what we had before.
* Refactor: make tests easier to read
* #1391: use enum for valid token types
This was previously only a description. This may seem like a breaking
change because OpenAPI would previously accept any string. However,
Joi also performs validation on this, so invalid values wouldn't work
previously either.
* #1391: Comment out default parameter values for now
This isn't the _right_ way, but it's the pragmatic solution. It's not
a big deal and this works as a stopgap solution.
* #1391: add todo note for api token schema fixes
* #1391: update snapshot
* Revert "#1391: modify patch schema to specify either object or array"
This reverts commit 0dd5d0faa1.
Turns out we need to allow much more than just objects and arrays.
I'll leave this as is for now.
* Chore(#1391): update comment explaining api token schema TODO
* #1391: modify some test code and add comment
* #1391: update tests and spec to allow 'any' type in schema
* Chore: remove comment
* #1391: add tests for context field stores
* #1391: add location header for public signup links
* #1391: fix query parameter description.
There was indeed a bug in the package, and this has been addressed
now, so we can go back to describing the params properly.
* Refactor(#1391): don't check for OpenAPI version in snapshot tests
remove the version from the received specification. The version
_is_ a required field, but having the version listed in automated
testing causes issues when trying to deploy new versions of the
API (due to mismatch between new tag versions etc). That's why we
remove it.
* #1391: test that the "semver" property is present
* #1391: test that the openapi version is valid semver
* Chore(#1391): comment formatting
* First version ready
* Final
* Refactor
* Update pat store
* Website revert
* Website revert
* Update
* Revert website
* Revert docs to main
* Revert docs to main
* Fix eslint
* Test
* Fix table name
* Fix(#1391): Expose API version even when not running via npm/yarn
## What
Expose the current Unleash version in the generated OpenAPI docs, even
when running via docker or other processes.
## Why
An OpenAPI spec without a version isn't valid. This causes some of our
generation tools (such as the one for documentation) to fail.
## How
By changing how we fetch the current version:
Previously, we used `process.env.npm_package_version!`. However, when
you're not running with yarn or npm, this is `undefined`. That causes
the version number to not be included when running tests and when
running using the official docker image.
Instead, we now use `version` from `lib/util/version`. This is the
same version as the one used by the UI config endpoint, so it should
be the same one as what the front end displays. To the best of my
knowledge, this _is_ the version of the API.
## Discussion
It _may_ be that I have misunderstood what the version represents, but
from what I can tell, it just exports what's listed as the version in
package.json.
The source code of the `lib/util/version` file is:
```ts
// export module version
require('pkginfo')(module, 'version');
const { version } = module.exports;
export default version;
module.exports = version;
```
* Refactor(#1391): rename imported variable for clarity
* PublicSignupTokens
* bug fix
* bug fixes and test
* bug fixes and test
* bug fixes and test
* Add feature flag
* tests
* tests
* Update 20220908093515-add-public-signup-tokens.js
Bug Fix
* task: use swc instead of ts-jest (#2042)
* Add a counter for total number of environments (#1964)
* add groupId to gradual rollout template (#2045)
* add groupId to gradual rollout template
* FMT
* Improve tabs UI on smaller devices (#2014)
* Improve tabs UI on smaller devices
* Improve tabs UI on smaller devices
* bug fix
* add proper scrollable tabs
* removed centered from Tabs (conflicts with scrollable)
* PR comments
* 4.15.0-beta.10
* Fix broken doc links (#2046)
## What
This PR fixes some broken links that have been hanging around in the
docs for what seems like a very long time.
## Why
As discovered by the link check in #1912, there are a fair few broken
links in the docs. Everyone hates broken links because it makes it
harder to understand what they were supposed to be pointing at.
## How
There are 3 types of links that have been fixed:
- Links that should have been internal but were absolute. E.g.
`https://docs.getunleash.io/path/article` that should have been
`./article.md`
- External links that have changed, such as Slack's API description
- GitHub links to files that either no longer exist or that have been
moved. These links generally pointed to `master`/`main`, meaning
they are subject to change. They have been replaced with permalinks
pointing to specific commits.
-----
* docs: fix slack api doc link
* docs: update links in migration guide
* docs: fix broken link to ancient feature schema validation
* docs: update links to v3 auth hooks
* docs: update broken link in the go sdk article
* Fix: use permalink for GitHub link
* docs: fix wrong google auth link
* 4.15.0
* 4.15.1
* docs: update link for symfony sdk (#2048)
The doc link appears to have pointed at an address that is no longer reachable. Instead, let's point to the equivalent GitHub link
Relates to and closes#2047
* docs: test broken links in website (#1912)
The action triggers manually as a first step to test this functionality. In the near future, we might schedule it
* Schedule link checker action (#2050)
Runs at 12:30 UTC on Mon, Tue, Wed, Thu and Fri
* fix: add env and project labels to feature updated metrics. (#2043)
* Revert workflow (#2051)
* update snapshot
* PR comments
* Added Events and tests
* Throw error if token not found
Co-authored-by: Christopher Kolstad <chriswk@getunleash.ai>
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
Co-authored-by: Gastón Fournier <gaston@getunleash.ai>
Co-authored-by: Ivar Conradi Østhus <ivarconr@gmail.com>
Co-authored-by: sjaanus <sellinjaanus@gmail.com>
* fix: varidate-edge-tokens-schema tokens type
This change fixes the description of the 'tokens' array of the
validate edge tokens schema.
The source of the error was a simple inverson of anyOf and items.
* fix: update snapshot
* feat: use unleash flags for embedded proxy
* feat: add a separate flag for the proxy frontend
* fix: setup unleash in dev
* fix: check flagResolver on each request
* fix: remove unleash client setup
* refactor: update frontend routes snapshot
* refactor: make batchMetrics flag dynamic
* fix: always check dynamic CORS origins config
* fix: make conditionalMiddleware work with the OpenAPI schema generation
Co-authored-by: olav <mail@olav.io>
* fix: use the frontend dir from the backend
* Build is now working
* Fix workflows
* Fix workflows
* Fix build PRs
* Test coverage workflow
* Test coverage
* Test coverage run
* Fix jest report
* refactor: add missing frontend build
* refactor: ignore frontend dir for coverage
* refactor: run frontend build in PRs
* refactor: run backend tests in PRs
* Revert "refactor: run backend tests in PRs"
This reverts commit 22cabddfd1.
* refactor: remove unused frontend build file
* refactor: test workflows in PR
* refactor: use a prepare script for the frontend
* refactor: simplify yarn build scripts
* refactor: fix check-release script
* Revert "refactor: test workflows in PR"
This reverts commit 496ae19404.
* refactor: remove unused gitignore lines
* refactor: remove renovate config from the frontend repo
* refactor: remove frontend repo license
* refactor: remove frontend repo changelog
* refactor: update frontend repo readme
* refactor: add frontend node_modules to dockerignore
* refactor: update the docker yarn.lock snapshot
Co-authored-by: olav <mail@olav.io>
* fix: remove unused exp flag
* fix: remove unused flag
* fix: add support for external flag resolver
* fix: rename flagsresolver to flagresolver
* fix: disable external flag resolver
* fix: refactor a bit
* fix: stop using unleash in server-dev
* fix: remove userGroups flag
* fix: revert bumping frontend
* refactor: remove unused API definition routes
* feat: embed proxy endpoints
* feat: check token metadata for alias if none is found
* fix: rename param
* feat: add test for retrieving token by alias
* fix: update schema
* fix: refactor to alias
* fix: refactor to null
* fix: update snapshot
* fix: update openapi snapshot
* fix: add check to getUserForToken
* refactor: add more token alias tests
* refactor: use timingSafeEqual for token comparisons
Co-authored-by: olav <mail@olav.io>
* refactor: batch feature metric inserts
* fix: use startOfHour when collapsing metrics
* refactor: avoid extra loop to sum yes/no values
* refactor: add experimental flag for batching metrics
* refactor: remove unused API definition routes
* feat: add support for proxy keys
* feat: support listening for any event
* feat: embed proxy endpoints
* refactor: add an experimental flag for the embedded proxy
## PR text
I realized that the tag changes we introduced in #1907 would be breaking for people who use the unleash-server package and implement their own endpoints on top of it (as we do in unleash-enterprise).
This change makes it possible to still use the old tags. However, these tags are purposefully not added to the root schema or the list of OpenAPI tag types. Any of our endpoints still using them (I think there is one in Unleash enterprise, see ivarconr/unleash-enterprise#109) should switch over when possible.
## Commits
* fix: Add deprecated openapi tags as valid tags for operations
* Docs: add explanatory comment for why the type exists.
* Docs: start experimenting with OpenAPI and docusaurus
* Docs: add docusaurus-theme-openapi-docs pkg
* Wip: current status
* Docs: Add 'docusaurus-plugin-api-docs'
* Move openapi into own sidebar; generate from localhost
* Chore: Update docusaurus plugin for OpenAPI
* Add website/yarn.lock to git
* Fix: fix CSS warning by using flex-end instead of end
* docs: make openapi generated code work again
* docs: make tags work properly with openapi sidebar
* Docs/chore: update OpenAPI tag scheme.
Add a whole bunch of new tags to make it easier to understand
available tags in OpenAPI.
* docs: point to new openapi docs from old api docs
* docs: typo
* Docs: link restructure
* docs: add operation indicators to openapi docs
* docs: change badge color for operations
* docs: update openapi-docs package
It now sorts tags the same as the schema
* docs: pluralize APIs in slug
* docs: update links to generated api docs
* docs: update openapi snapshot tests with new tags
* docs: conditionally load spec from localhost or from file
* docs: Remove changes relating to immediate switchover
* refactor: rename types; extract into separate file
* docs: fix api doc links
The hostname strategy will not work correctly with the playground because it depends on external state. In its constructor, it tries to query the environment or use the os.hostname function to determine what its current hostname is. This means that no matter what the user does in the playground, they can’t affect the results of this strategy. It’s also unlikely that it will be true. And if it is, it probably won’t be true for their clients.
In theory, we could accept a hostname property on the Unleash context and use the provided hostname in the address. However, I’m afraid that it’ll make users think that they can impact the hostname strategy by setting the property on their context, when that doesn’t do anything outside of the playground. It would also make the playground evaluate things differently from a regular SDK and I’m not sure that that’s something we want.
Instead, this change to the API makes the feature evaluate to 'unknown' or `false` (depending on constraints).
* Feat: return reasons why a feature evaluated to true or false
Note: this is very rough and just straight ripped from the nodejs
client. It will need a lot of work, but is a good place to start
* Feat: add suggested shape for new payload
* Chore: minor cleanup
* Wip: make server compile again
* Remove unused schema ref
* Export new schemas
* Chore: fix some tests to use sub property
* Fix: fix some tests
* Refactor: rename some variables, uncomment some stuff
* Add segments type to bootstrap options
* Add segments capability to offline feature evaluator
* Fix function calls after turning params into an option abject
* Feat: test strategy order, etc
* Feat: add test to check that all strats are returned correctly
* Feat: allow you to include strategy ids in clients
* Wip: hook up segments in the offline client.
Note: compared to regular clients, they still fail
* Feat: add segments validation
* Fix: fix test case invariant.
* Chore: revert to returning only `boolean` from strategies.
This _should_ make it work with custom strategies too 🤞
* Feat: make more properties of the returned feature required
* Wip: add some comments and unfinished tests for edge cases
* Feat: add `isEnabledInCurrentEnvironment` prop
* Feat: consider more strategy failure cases
* Feat: test that isenabledinenvironment matches expectations
* Feat: add unknown strategies
* Fix: fix property access typo
* Feat: add unknown strategy for fallback purposes
* Feat: test edge case: all unknown strategies
* Feat: add custom strategy to arbitrary
* Feat: test that features can be true, even if not enabled in env
* Chore: add some comments
* Wip: fix sdk tests
* Remove comments, improve test logging
* Feat: add descriptions and examples to playground feature schema
* Switch `examples` for `example`
* Update schemas with descriptions and examples
* Fix: update snapshot
* Fix: openapi example
* Fix: merge issues
* Fix: fix issue where feature evaluation state was wrong
* Chore: update openapi spec
* Fix: fix broken offline client tests
* Refactor: move schemas into separate files
* Refactor: remove "reason" for incomplete evaluation.
The only instances where evaluation is incomplete is when we don't
know what the strategy is.
* Refactor: move unleash node client into test and dev dependencies
* Wip: further removal of stuff
* Chore: remove a bunch of code that we don't use
* Chore: remove comment
* Chore: remove unused code
* Fix: fix some prettier errors
* Type parameters in strategies to avoid `any`
* Fix: remove commented out code
* Feat: make `id` required on playground strategies
* Chore: remove redundant type
* Fix: remove redundant if and fix fallback evaluation
* Refactor: reduce nesting and remove duplication
* Fix: remove unused helper function
* Refactor: type `parameters` as `unknown`
* Chore: remove redundant comment
* Refactor: move constraint code into a separate file
* Refactor: rename `unleash` -> `feature-evaluator`
* Rename class `Unleash` -> `FeatureEvaluator`
* Refactor: remove this.ready and sync logic from feature evaluator
* Refactor: remove unused code, rename config type
* Refactor: remove event emission from the Unleash client
* Remove unlistened-for events in feature evaluator
* Refactor: make offline client synchronous; remove code
* Fix: update openapi snapshot after adding required strategy ids
* Feat: change `strategies` format.
This commit changes the format of a playground feature's `strategies`
properties from a list of strategies to an object with properties
`result` and `data`. It looks a bit like this:
```ts
type Strategies = {
result: boolean | "unknown",
data: Strategy[]
}
```
The reason is that this allows us to avoid the breaking change that
was previously suggested in the PR:
`feature.isEnabled` used to be a straight boolean. Then, when we found
out we couldn't necessarily evaluate all strategies (custom strats are
hard!) we changed it to `boolean | 'unevaluated'`. However, this is
confusing on a few levels as the playground results are no longer the
same as the SDK would be, nor are they strictly boolean anymore.
This change reverts the `isEnabled` functionality to what it was
before (so it's always a mirror of what the SDK would show).
The equivalent of `feature.isEnabled === 'unevaluated'` now becomes
`feature.isEnabled && strategy.result === 'unknown'`.
* Fix: Fold long string descriptions over multiple lines.
* Fix: update snapshot after adding line breaks to descriptions
