## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
Adds projects user and group -usage information to the dialog shown when
user wants to delete a project role
<img width="670" alt="Skjermbilde 2023-08-10 kl 08 28 40"
src="https://github.com/Unleash/unleash/assets/707867/a1df961b-2d0f-419d-b9bf-fedef896a84e">
---------
Co-authored-by: Nuno Góis <github@nunogois.com>
Unless you set `additionalProperties` to `false`, additional properties
are always
allowed. By explicitly setting it to `true` the generated
examples contain additional properties, e.g.:
```json
{
"tokens": [
"aproject:development.randomstring",
"[]:production.randomstring"
],
"additionalProp1": {}
}
```
By removing the explicit annotation, we still allow additional
properties, but we don't get them in examples:
```json
{
"tokens": [
"aproject:development.randomstring",
"[]:production.randomstring"
]
}
```
### What
This PR changes the slack-app addon to use slack-api's scheduleMessage
instead of postMessage.
### Why
When using postMessage we had to find the channel id in order to be able
to post the message to the channel. scheduleMessage allows using the
channel name instead of the id, which saves the entire struggle of
finding the channel name. It did mean that we had to move to defining
blocks of content instead of the easier formatting we did with
postMessage.
### Message look
https://linear.app/unleash/issue/2-1136/custom-root-roles-documentation
- [Adds documentation referencing custom root
roles](https://unleash-docs-git-docs-custom-root-roles-unleash-team.vercel.app/reference/rbac);
- [Adds a "How to create and assign custom root roles" how-to
guide](https://unleash-docs-git-docs-custom-root-roles-unleash-team.vercel.app/how-to/how-to-create-and-assign-custom-root-roles);
- Standardizes "global" roles to "root" roles;
- Standardizes "standard" roles to "predefined" roles to better reflect
their behavior and what is shown in our UI;
- Updates predefined role descriptions and makes them consistent;
- Updates the side panel description of the user form;
- Includes some boy scouting with some tiny fixes of things identified
along the way (e.g. the role form was persisting old data when closed
and re-opened);
Questions:
- Is it worth expanding the "Assigning custom root roles" section in the
"How to create and assign custom root roles" guide to include the steps
for assigning a root role for each entity (user, service account,
group)?
- Should this PR include an update to the existing "How to create and
assign custom project roles" guide? We've since updated the UI;
---------
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
## About the changes
Open API is creating 2 resources under the same URL path, we want them
aggregated:
```shell
$ curl -s https://us.app.unleash-hosted.com/ushosted/docs/openapi.json | jq '.paths."/api/admin/strategies/{name}" | keys'
[
"delete",
"get"
]
gaston@gaston-Summit-E16Flip-A12UCT: ~/poc/terraform-provider-unleash on main [!$]
$ curl -s https://us.app.unleash-hosted.com/ushosted/docs/openapi.json | jq '.paths."/api/admin/strategies/{strategyName}" | keys'
[
"put"
]
```
Note one is under `{name}` while the other is under `{strategyName}`
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
Fixes an issue where project role deletion validation didn't validate
against project roles being connected to groups
We've been struggling with not getting hold of all channels. Reading the
documentation I found the support for next_cursor in the
response_metadata, as long as the response has a next_cursor, there are
responses left, so this PR adds a loop for querying until the next
cursor or the channels list is undefined or `''`. This should solve the
issue Wayfair had as well.
https://linear.app/unleash/issue/2-1311/add-a-new-prometheus-metric-with-custom-root-roles-in-use
As a follow-up to https://github.com/Unleash/unleash/pull/4435, this PR
adds a metric for total custom root roles in use by at least one entity:
users, service accounts, groups.
`custom_root_roles_in_use_total`
Output from `http://localhost:4242/internal-backstage/prometheus`:
```
# HELP process_cpu_user_seconds_total Total user CPU time spent in seconds.
# TYPE process_cpu_user_seconds_total counter
process_cpu_user_seconds_total 0.060755
# HELP process_cpu_system_seconds_total Total system CPU time spent in seconds.
# TYPE process_cpu_system_seconds_total counter
process_cpu_system_seconds_total 0.01666
# HELP process_cpu_seconds_total Total user and system CPU time spent in seconds.
# TYPE process_cpu_seconds_total counter
process_cpu_seconds_total 0.077415
# HELP process_start_time_seconds Start time of the process since unix epoch in seconds.
# TYPE process_start_time_seconds gauge
process_start_time_seconds 1691420275
# HELP process_resident_memory_bytes Resident memory size in bytes.
# TYPE process_resident_memory_bytes gauge
process_resident_memory_bytes 199196672
# HELP nodejs_eventloop_lag_seconds Lag of event loop in seconds.
# TYPE nodejs_eventloop_lag_seconds gauge
nodejs_eventloop_lag_seconds 0
# HELP nodejs_eventloop_lag_min_seconds The minimum recorded event loop delay.
# TYPE nodejs_eventloop_lag_min_seconds gauge
nodejs_eventloop_lag_min_seconds 0.009076736
# HELP nodejs_eventloop_lag_max_seconds The maximum recorded event loop delay.
# TYPE nodejs_eventloop_lag_max_seconds gauge
nodejs_eventloop_lag_max_seconds 0.037683199
# HELP nodejs_eventloop_lag_mean_seconds The mean of the recorded event loop delays.
# TYPE nodejs_eventloop_lag_mean_seconds gauge
nodejs_eventloop_lag_mean_seconds 0.011063251638989169
# HELP nodejs_eventloop_lag_stddev_seconds The standard deviation of the recorded event loop delays.
# TYPE nodejs_eventloop_lag_stddev_seconds gauge
nodejs_eventloop_lag_stddev_seconds 0.0013618102764025837
# HELP nodejs_eventloop_lag_p50_seconds The 50th percentile of the recorded event loop delays.
# TYPE nodejs_eventloop_lag_p50_seconds gauge
nodejs_eventloop_lag_p50_seconds 0.011051007
# HELP nodejs_eventloop_lag_p90_seconds The 90th percentile of the recorded event loop delays.
# TYPE nodejs_eventloop_lag_p90_seconds gauge
nodejs_eventloop_lag_p90_seconds 0.011321343
# HELP nodejs_eventloop_lag_p99_seconds The 99th percentile of the recorded event loop delays.
# TYPE nodejs_eventloop_lag_p99_seconds gauge
nodejs_eventloop_lag_p99_seconds 0.013688831
# HELP nodejs_active_resources Number of active resources that are currently keeping the event loop alive, grouped by async resource type.
# TYPE nodejs_active_resources gauge
nodejs_active_resources{type="FSReqCallback"} 1
nodejs_active_resources{type="TTYWrap"} 3
nodejs_active_resources{type="TCPSocketWrap"} 5
nodejs_active_resources{type="TCPServerWrap"} 1
nodejs_active_resources{type="Timeout"} 1
nodejs_active_resources{type="Immediate"} 1
# HELP nodejs_active_resources_total Total number of active resources.
# TYPE nodejs_active_resources_total gauge
nodejs_active_resources_total 12
# HELP nodejs_active_handles Number of active libuv handles grouped by handle type. Every handle type is C++ class name.
# TYPE nodejs_active_handles gauge
nodejs_active_handles{type="WriteStream"} 2
nodejs_active_handles{type="ReadStream"} 1
nodejs_active_handles{type="Socket"} 5
nodejs_active_handles{type="Server"} 1
# HELP nodejs_active_handles_total Total number of active handles.
# TYPE nodejs_active_handles_total gauge
nodejs_active_handles_total 9
# HELP nodejs_active_requests Number of active libuv requests grouped by request type. Every request type is C++ class name.
# TYPE nodejs_active_requests gauge
nodejs_active_requests{type="FSReqCallback"} 1
# HELP nodejs_active_requests_total Total number of active requests.
# TYPE nodejs_active_requests_total gauge
nodejs_active_requests_total 1
# HELP nodejs_heap_size_total_bytes Process heap size from Node.js in bytes.
# TYPE nodejs_heap_size_total_bytes gauge
nodejs_heap_size_total_bytes 118587392
# HELP nodejs_heap_size_used_bytes Process heap size used from Node.js in bytes.
# TYPE nodejs_heap_size_used_bytes gauge
nodejs_heap_size_used_bytes 89642552
# HELP nodejs_external_memory_bytes Node.js external memory size in bytes.
# TYPE nodejs_external_memory_bytes gauge
nodejs_external_memory_bytes 1601594
# HELP nodejs_heap_space_size_total_bytes Process heap space size total from Node.js in bytes.
# TYPE nodejs_heap_space_size_total_bytes gauge
nodejs_heap_space_size_total_bytes{space="read_only"} 0
nodejs_heap_space_size_total_bytes{space="old"} 70139904
nodejs_heap_space_size_total_bytes{space="code"} 3588096
nodejs_heap_space_size_total_bytes{space="map"} 2899968
nodejs_heap_space_size_total_bytes{space="large_object"} 7258112
nodejs_heap_space_size_total_bytes{space="code_large_object"} 1146880
nodejs_heap_space_size_total_bytes{space="new_large_object"} 0
nodejs_heap_space_size_total_bytes{space="new"} 33554432
# HELP nodejs_heap_space_size_used_bytes Process heap space size used from Node.js in bytes.
# TYPE nodejs_heap_space_size_used_bytes gauge
nodejs_heap_space_size_used_bytes{space="read_only"} 0
nodejs_heap_space_size_used_bytes{space="old"} 66992120
nodejs_heap_space_size_used_bytes{space="code"} 2892640
nodejs_heap_space_size_used_bytes{space="map"} 2519280
nodejs_heap_space_size_used_bytes{space="large_object"} 7026824
nodejs_heap_space_size_used_bytes{space="code_large_object"} 983200
nodejs_heap_space_size_used_bytes{space="new_large_object"} 0
nodejs_heap_space_size_used_bytes{space="new"} 9236136
# HELP nodejs_heap_space_size_available_bytes Process heap space size available from Node.js in bytes.
# TYPE nodejs_heap_space_size_available_bytes gauge
nodejs_heap_space_size_available_bytes{space="read_only"} 0
nodejs_heap_space_size_available_bytes{space="old"} 1898360
nodejs_heap_space_size_available_bytes{space="code"} 7328
nodejs_heap_space_size_available_bytes{space="map"} 327888
nodejs_heap_space_size_available_bytes{space="large_object"} 0
nodejs_heap_space_size_available_bytes{space="code_large_object"} 0
nodejs_heap_space_size_available_bytes{space="new_large_object"} 16495616
nodejs_heap_space_size_available_bytes{space="new"} 7259480
# HELP nodejs_version_info Node.js version info.
# TYPE nodejs_version_info gauge
nodejs_version_info{version="v18.16.0",major="18",minor="16",patch="0"} 1
# HELP nodejs_gc_duration_seconds Garbage collection duration by kind, one of major, minor, incremental or weakcb.
# TYPE nodejs_gc_duration_seconds histogram
# HELP http_request_duration_milliseconds App response time
# TYPE http_request_duration_milliseconds summary
# HELP db_query_duration_seconds DB query duration time
# TYPE db_query_duration_seconds summary
db_query_duration_seconds{quantile="0.1",store="api-tokens",action="getAllActive"} 0.03091475
db_query_duration_seconds{quantile="0.5",store="api-tokens",action="getAllActive"} 0.03091475
db_query_duration_seconds{quantile="0.9",store="api-tokens",action="getAllActive"} 0.03091475
db_query_duration_seconds{quantile="0.95",store="api-tokens",action="getAllActive"} 0.03091475
db_query_duration_seconds{quantile="0.99",store="api-tokens",action="getAllActive"} 0.03091475
db_query_duration_seconds_sum{store="api-tokens",action="getAllActive"} 0.03091475
db_query_duration_seconds_count{store="api-tokens",action="getAllActive"} 1
# HELP feature_toggle_update_total Number of times a toggle has been updated. Environment label would be "n/a" when it is not available, e.g. when a feature toggle is created.
# TYPE feature_toggle_update_total counter
# HELP feature_toggle_usage_total Number of times a feature toggle has been used
# TYPE feature_toggle_usage_total counter
# HELP feature_toggles_total Number of feature toggles
# TYPE feature_toggles_total gauge
feature_toggles_total{version="5.3.0"} 31
# HELP users_total Number of users
# TYPE users_total gauge
users_total 1011
# HELP projects_total Number of projects
# TYPE projects_total gauge
projects_total 4
# HELP environments_total Number of environments
# TYPE environments_total gauge
environments_total 10
# HELP groups_total Number of groups
# TYPE groups_total gauge
groups_total 5
# HELP roles_total Number of roles
# TYPE roles_total gauge
roles_total 11
# HELP custom_root_roles_total Number of custom root roles
# TYPE custom_root_roles_total gauge
custom_root_roles_total 3
# HELP custom_root_roles_in_use_total Number of custom root roles in use
# TYPE custom_root_roles_in_use_total gauge
custom_root_roles_in_use_total 2
# HELP segments_total Number of segments
# TYPE segments_total gauge
segments_total 5
# HELP context_total Number of context
# TYPE context_total gauge
context_total 7
# HELP strategies_total Number of strategies
# TYPE strategies_total gauge
strategies_total 5
# HELP client_apps_total Number of registered client apps aggregated by range by last seen
# TYPE client_apps_total gauge
client_apps_total{range="allTime"} 0
client_apps_total{range="30d"} 0
client_apps_total{range="7d"} 0
# HELP saml_enabled Whether SAML is enabled
# TYPE saml_enabled gauge
saml_enabled 1
# HELP oidc_enabled Whether OIDC is enabled
# TYPE oidc_enabled gauge
oidc_enabled 0
# HELP client_sdk_versions Which sdk versions are being used
# TYPE client_sdk_versions counter
# HELP optimal_304_diffing Count the Optimal 304 diffing with status
# TYPE optimal_304_diffing counter
# HELP db_pool_min Minimum DB pool size
# TYPE db_pool_min gauge
db_pool_min 0
# HELP db_pool_max Maximum DB pool size
# TYPE db_pool_max gauge
db_pool_max 4
# HELP db_pool_free Current free connections in DB pool
# TYPE db_pool_free gauge
db_pool_free 0
# HELP db_pool_used Current connections in use in DB pool
# TYPE db_pool_used gauge
db_pool_used 4
# HELP db_pool_pending_creates how many asynchronous create calls are running in DB pool
# TYPE db_pool_pending_creates gauge
db_pool_pending_creates 0
# HELP db_pool_pending_acquires how many acquires are waiting for a resource to be released in DB pool
# TYPE db_pool_pending_acquires gauge
db_pool_pending_acquires 24
```
This PR stabilizes the playground and feature types endpoints by
changing their tag from 'Unstable' to respectively 'Playground' and
'Feature Types'. It also moves the existing feature type endpoint (which
was previously tagged as 'Features') to the 'Feature Types' tag.
In a new fresh Unleash instance with cache enabled this can cause
feature toggles to never get updated.
We saw in our client that the ETag was ETag: "60e35fba:null" Which
looked incorrect for us.
I also did manual testing and if the andWhere had a value of largerThan
higher than whatever the id was then we would get back { max: null }.
This should fix that issue.
This PR adds an e2e test to the OpenAPI tests that checks that all
openapi operations have both summaries and descriptions. It also fixes
the few schemas that were missing one or the other.
Enable strict schema validation by default. It can still be overridden
by explicitly setting it to false.
I've also fixed the validation errors that appeared when turning it on.
I've opted for the simplest route and changed the schemas to comply with
the tests.
## About the changes
We are losing some events because of not having "created by" which is
required by a constraint in the DB. One scenario where this can happen
is with the default user admin, because it doesn't have a username or an
email (unless configured by the administrator).
Our code makes assumptions on the existence of one of these 2 attributes
(e.g.
248118af7c/src/lib/services/user-service.ts (L220-L222)).
Event lost metrics:
## Discussion points
The solution proposed here is falling back to a default value. I've
chosen `"admin"` because it covers one of the use cases, but it can also
be `"system"` mimicking
248118af7c/src/lib/services/user-service.ts (L32)
which is used as a default, or `"unknown"` which is sometimes used as a
default:
248118af7c/src/lib/services/project-service.ts (L57)
Anyway, I believe it's better not to lose the event rather than be
accurate with the "created by" that can be fixed later
This PR adds potentially stale events as available to all addons and
adds a formatted message.
## Discussion
I'd still be interested in hearing input on whether the event type
should be selectable by addons or not before we start emitting the
event. I'm leaning towards "yes", but I'll take your thoughts into
consideration.
This PR does **one** thing:
it changes the events for potentially stale to:
- Only being emitted when potentially stale gets turned on
- In doing so, it also simplifies the event that's getting emitted,
removing the `data` property.
- The event is also renamed to better match the existing
`feature-stale-on` and `...-off` events.
The addon listening was broken out into a separate PR (#4279)
## Old description
This change lets all addons listen for events when features get marked
or unmarked as potentially stale.
### Discussion
#### All addons?
Should this be available to all addons? I can't see a reason why it
shouldn't be available to all addons, but I might be missing
something.
**Update**: spoke to a couple people. Can see no reason why this isn't
okay.
#### Should it be behind a flag?
The feature is still behind a flag, but the event type is not. Should
we gate the event being available until we actually emit the event?
That would require some more code, but could yield less potential
confusion.
Open to hearing your thoughts.
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
wrap reorder event creation to strategy variant feature
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
<!-- Does it close an issue? Multiple? -->
Closes #
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
## Discussion points
<!-- Anything about the PR you'd like to discuss before it gets merged?
Got any questions or doubts? -->
---------
Signed-off-by: andreas-unleash <andreas@getunleash.ai>
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
Fixes a bug around createStrategy
Fix: Create/Store strategy reorder event only when feature is on
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
<!-- Does it close an issue? Multiple? -->
Closes #
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
## Discussion points
<!-- Anything about the PR you'd like to discuss before it gets merged?
Got any questions or doubts? -->
Signed-off-by: andreas-unleash <andreas@getunleash.ai>
This PR updates the feature type service by adding a new
`updateLifetime` method. This method handles the connection between the
API (#4256) and the store (#4252).
I've also added some new e2e tests to ensure that the API behaves as
expected.
This PR adds an operation and accompanying openapi docs for the new
"update feature type lifetime" API operation.
It also fixes an oversight where the other endpoint on the same
controller didn't use `respondWithValidation`.
Note: the API here is a suggestion. I'd like to hear whether you agree
with this implementation or not.
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
When reordering strategies for a feature environment:
- Adds stop when CR are enabled
- Emits an event
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
<!-- Does it close an issue? Multiple? -->
Closes #
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
## Discussion points
<!-- Anything about the PR you'd like to discuss before it gets merged?
Got any questions or doubts? -->
---------
Signed-off-by: andreas-unleash <andreas@getunleash.ai>
This PR activates the event emission that was prepared for in
https://github.com/Unleash/unleash/pull/4239.
It emits events (behind a flag) when something is marked as potentially
stale or the opposite.
It takes the features returned from the store and creates events out of
them.
The events only contain data, no preData. This is because the preData
can easily be inferred and because it gives a nicer event in the event
log.
Here is an image of the difference. The top event uses only data, so it
shows the name of the feature and the new potentiallyStale status. The
bottom event uses both preData and data, so it only shows the new
potentiallyStale status and not the feature name (unless you show the
raw event):
Should not be merged before #4239. Merge that and then rebase this off
main or cherry the commit.
## Discussion
### `preData`
Should we also use preData or is it enough to use only data? It seems
unnecessary in this event, but I'm open to hearing your thoughts.
### event author: `createdBy`
I've set `unleash-system` as the `createdBy` property on these events
because they are generated by the system. I found the same string used
some other places. However, it may be that there we want to use a
different author.
This PR adds updates the potentially stale status change events whenever
the potentially stale update function is run.
No events are emitted yet. While the emission is only a few lines of
code, I'd like to do that in a separate PR so that we can give it the
attention it deserves in the form of tests, etc.
This PR also moves the potentially stale update functionality from the
`update` method to only being done in the
`updatePotentiallyStaleFeatures` method. This keeps all functionality
related to marking `potentiallyStale` in one place.
The emission implementation was removed in
4fb7cbde03
## The update queries
While it would be possible to do the state updates in a single query
instead of three separate ones, wrangling this into knex proved to be
troublesome (and would also probably be harder to understand and reason
about). The current solution uses three smaller queries (one select, two
updates), as Jaanus suggested in a private slack thread.
This reverts commit 16e3799b9a.
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
<!-- Does it close an issue? Multiple? -->
Closes #
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
## Discussion points
<!-- Anything about the PR you'd like to discuss before it gets merged?
Got any questions or doubts? -->
https://linear.app/unleash/issue/2-1232/implement-first-iteration-of-the-new-slack-app-addon
This PR implements the first iteration of the new Slack App addon.
Unlike the old Slack addon, this one uses a Slack App (bot) that is
installed to Slack workspaces in order to post messages. This uses
`@slack/web-api`, which internally uses the latest Slack API endpoints
like `postMessage`.
This is currently behind a flag: `slackAppAddon`.
The current flow is that the Unleash Slack App is installed from
whatever source:
- Unleash addons page;
- Direct link;
- https://unleash-slack-app.vercel.app/ (temporary URL);
- Slack App Directory (in the future);
- Etc;
After installed, we resolve the authorization to an `access_token` that
the user can paste into the Unleash Slack App addon configuration form.
https://github.com/Unleash/unleash/assets/14320932/6a6621b9-5b8a-4921-a279-30668be6d46c
Co-authored by: @daveleek
---------
Co-authored-by: David Leek <david@getunleash.io>
This PR lays most of the groundwork required for emitting events when
features are marked as potentially stale by Unleash. It does **not**
emit any events just yet. The summary is:
- periodically look for features that are potentially stale and mark
them (set to run every 10 seconds for now; can be changed)
- when features are updated, if the update data contains changes to the
feature's type or createdAt date, also update the potentially stale
status.
It is currently about 220 lines of tests and about 100 lines of
application code (primarily db migration and two new methods on the
IFeatureToggleStore interface).
The reason I wanted to put this into a single PR (instead of just the db
migration, then just the potentially stale marking, then the update
logic) is:
If users get the db migration first, but not the rest of the update
logic until the events are fired, then they could get a bunch of new
events for features that should have been marked as potentially stale
several days/weeks/months ago. That seemed undesirable to me, so I
decided to bunch those changes together. Of course, I'd be happy to
break it into smaller parts.
## Rules
A toggle will be marked as potentially stale iff:
- it is not already stale
- its createdAt date is older than its feature type's expected lifetime
would dictate
## Migration
The migration adds a new `potentially_stale` column to the features
table and sets this to true for any toggles that have exceeded their
expected lifetime and that have not already been marked as `stale`.
## Discussion
### The `currentTime` parameter of `markPotentiallyStaleFeatures`
The `markPotentiallyStaleFetaures` method takes an optional
`currentTime` parameter. This was added to make it easier to test (so
you can test "into the future"), but it's not used in the application.
We can rewrite the tests to instead update feature toggles manually, but
that wouldn't test the actual marking method. Happy to discuss.
## About the changes
Fix un-awaited promise on batch variant update - reduce function allowed
TS to skip Promise type.
---------
Co-authored-by: Gastón Fournier <gaston@getunleash.io>
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
Wraps the whole `registerClientMetrics` function with try/catch to
return 400 on error
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
<!-- Does it close an issue? Multiple? -->
Closes #
[1-1037](https://linear.app/unleash/issue/1-1037/return-4xx-error-for-incorrect-metrics-input)
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
## Discussion points
<!-- Anything about the PR you'd like to discuss before it gets merged?
Got any questions or doubts? -->
---------
Signed-off-by: andreas-unleash <andreas@getunleash.ai>
## About the changes
- Adding descriptions and examples to tag and tag types schemas
- Adding standard errors, summaries, and descriptions to tag and tag
types endpoints
- Some improvements on compilation errors
---------
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
Make each error class have to define its own status code. This makes
it easier for developers to see which code an error corresponds to and
means less jumping back and forth between files. In other words:
improved locality.
Unfortunately, the long switch needs to stay in case we get errors
thrown that aren't of the Unleash Error type, but we can move it to
the `fromLegacyError` file instead.
Tradeoff analysis by @kwasniew:
+ I like the locality of error to code reasoning
- now HTTP leaks to the non-HTTP code that throws those errors e.g. application services
If we had other delivery mechanisms other than HTTP then it wouldn't make sense to couple error codes to one protocol (HTTP). But since we're mostly doing web it may not be a problem.
@thomasheartman's response:
This is a good point and something I hadn't considered. The same data was always available on those errors (by using the same property), I've just made the declaration local to each error instead of something that the parent class handles. The idea was to make it easier to create new error classes with their corresponding error codes. Because the errors are intended to be API errors (or at least, I've always considered them to be that), I think that makes sense.
Taking your comment into consideration, I still think it's the right thing to do, but I'm not bullish about it. We could always walk it back later if we find that it's not appropriate. The old code is still available and we could easily enough roll back this change if we find that we want to decouple it later.
## About the changes
Include a new configuration parameter to be able to specify
source_type_name. This is an opt-in feature which provides backward
compatibility to our existing users.
Closes#4109
## Discussion points
Maybe this should be hardcoded to `Unleash` but this gives additional
flexibility
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
Adds description and summary to `favorite` endpoints
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
<!-- Does it close an issue? Multiple? -->
Closes #
[1-1098](https://linear.app/unleash/issue/1-1098/openapi-features-favorites)
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
## Discussion points
<!-- Anything about the PR you'd like to discuss before it gets merged?
Got any questions or doubts? -->
---------
Signed-off-by: andreas-unleash <andreas@getunleash.ai>
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
In some of the places we used `NoAccessError` for permissions, other
places we used it for a more generic 403 error with a different
message. This refactoring splits the error type into two distinct
types instead to make the error messages more consistent.
### What
Adds a quick and dirty description to requestPerSeconds and
segmentedRequestPerSecondsSchema so the enterprise /rps endpoint has
better API docs.
---------
Co-authored-by: Simon Hornby <liquidwicked64@gmail.com>
This PR fixes an issue where events generated during a db transaction
would get published before the transaction was complete. This caused
errors in some of our services that expected the data to be stored
before the transaction had been commited. Refer to [linear issue
1-1049](https://linear.app/unleash/issue/1-1049/event-emitter-should-emit-events-after-db-transaction-is-commited-not)
for more info.
Fixes 1-1049.
## Changes
The most important change here is that the `eventStore` no longer emits
events when they happen (because that can be in the middle of a
transaction). Instead, events are stored with a new `announced` column.
The new event announcer service runs on a schedule (every second) and
publishes any new events that have not been published.
Parts of the code have largely been lifted from the
`client-application-store`, which uses a similar logic.
I have kept the emitting of the event within the event store because a
lot of other services listen to events from this store, so removing that
would require a large rewrite. It's something we could look into down
the line, but it seems like too much of a change to do right now.
## Discussion
### Terminology:
Published vs announced? We should settle on one or the other. Announced
is consistent with the client-application store, but published sounds
more fitting for events.
### Publishing and marking events as published
The current implementation fetches all events that haven't been marked
as announced, sets them as announced, and then emits them. It's possible
that Unleash would crash in the interim or something else might happen,
causing the events not to get published. Maybe it would make sense to
just fetch the events and only mark them as published after the
announcement? On the other hand, that might get us into other problems.
Any thoughts on this would be much appreciated.
## About the changes
Fix OpenAPI definitions for endpoint
`/api/admin/projects/{projectId}/features/{featureName}/environments/{environment}`
and similar.
Apparently Jest doesn't like it when you use multiple stringContaining
statements for one property. Only the last `stringContaining`
statement is evaluated while the others are ignored.
This means that a lot of these assertions were never checked. To fix
that, I've extracted them into separate assertions.
### What
This PR adds documentation for our endpoints that are covered by our
"Events" tag. It also adds a type for all valid events, and then uses
this as valid values for type argument.
This PR updates endpoints and schemas for the API tokens tag.
As part of that, they also handle oneOf openapi validation errors and improve the console output for the enforcer tests.
Background (general): I don’t think it creates any value. I don’t think
people check the snapshot explicitly (especially now that it’s so large
that it gets hidden by default), and it only creates an extra chore that
you forget to do all the time (at least I do).
Background (why now?): I’ve been fighting a weird issue with the
snapshot test for the [API tokens OpenAPI
PR](https://github.com/Unleash/unleash/pull/4137). The tests run fine
locally, but fail in CI. The tests seem to be using an old version of
the snapshot for the test. However, when I check the snapshot in the
files tab, it’s got the new and correct values. I’ve spent probably more
than an hour trying to fix and debug this today but to no avail.
###What
Adds an optional sensitive parameter for customHeaders to all current
addons. It is sensitive because the user might be including api key headers.
## What
This adds openapi documentation for the Auth tagged operations and
connected schemas.
## Discussion points
Our user schema seems to be exposing quite a bit of internal fields, I
flagged the isApi field as deprecated, I can imagine quite a few of
these fields also being deprecated to prepare for removal in next major
version, but I was unsure which ones were safe to do so with.
## Observation
We have some technical debt around the shape of the schema we're
claiming we're returning and what we actually are returning. I believe
@gastonfournier also observed this when we turned on validation for our
endpoints.
---------
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
This removes the experimental feature flag that defaulted to turn off
telemetry collection
Catch cases where the segment doesn't exist and populate that error
message with more info: it now says that a segment
with <id> doesn't exist instead of just 'No row'.
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
Adds a UI that shows current status of version and feature usage
collection configuration, and a presence in the configuration menu +
menu bar.
Configuring these features is done by setting environment variables. The
version info collection is an existing feature that we're making more
visible, the feature usage collection feature is a new feature that has
it's own environment configuration but also depends on version info
collection being active to work.
When version collection is turned off and the experimental feature flag
for feature usage collection is turned off:
<img width="1269" alt="image"
src="https://github.com/Unleash/unleash/assets/707867/435a07da-d238-4b5b-a150-07e3bd6b816f">
When version collection is turned on and the experimental feature flag
is off:
<img width="1249" alt="image"
src="https://github.com/Unleash/unleash/assets/707867/8d1a76c5-99c9-4551-9a4f-86d477bbbf6f">
When the experimental feature flag is enabled, and version+telemetry is
turned off:
<img width="1239" alt="image"
src="https://github.com/Unleash/unleash/assets/707867/e0bc532b-be94-4076-bee1-faef9bc48a5b">
When version collection is turned on, the experimental feature flag is
enabled, and telemetry collection is turned off:
<img width="1234" alt="image"
src="https://github.com/Unleash/unleash/assets/707867/1bd190c1-08fe-4402-bde3-56f863a33289">
When version collection is turned on, the experimental feature flag is
enabled, and telemetry collection is turned on:
<img width="1229" alt="image"
src="https://github.com/Unleash/unleash/assets/707867/848912cd-30bd-43cf-9b81-c58a4cbad1e4">
When version collection is turned off, the experimental feature flag is
enabled, and telemetry collection is turned on:
<img width="1241" alt="image"
src="https://github.com/Unleash/unleash/assets/707867/d2b981f2-033f-4fae-a115-f93e0653729b">
---------
Co-authored-by: sighphyre <liquidwicked64@gmail.com>
Co-authored-by: Nuno Góis <github@nunogois.com>
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
### What
Added documentation for context fields and endpoints tagged with the
"Context" tag.
---------
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
This PR adds strategy titles as an optional bit of data added to client
features. It's only added when prompted.
## Discussion points:
### getPlaygroundFeatures
The optional `includeStrategyId` parameter has been replaced by a
`getPlaygroundFeatures` in the service (and in the underlying store).
The playground was the only place that used this specific include, so
instead of adding more and making the interface for that method more
complex, I created a new method that deals specifically with the
playground.
The underlying store still uses an `optionalIncludes` parameter,
however. I have a plan to make that interface more fluid, but I'd like
to propose that in a follow-up PR.
As part of the move to a unified domain this PR updates the default
EMAIL_SENDER to noreply@getunleash.io . Should not be merged/deployed
until we've verified DMARC, DKIM for the new domain.
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
Remove strategy improvements flag
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
<!-- Does it close an issue? Multiple? -->
Closes #
[1-1048](https://linear.app/unleash/issue/1-1048/remove-strategyimprovements-flag)
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
## Discussion points
<!-- Anything about the PR you'd like to discuss before it gets merged?
Got any questions or doubts? -->
---------
Signed-off-by: andreas-unleash <andreas@getunleash.ai>
## About the changes
This makes these permissions not available for selection. In particular
`UPDATE_ROLE`, `CREATE_API_TOKEN`, `UPDATE_API_TOKEN`,
`DELETE_API_TOKEN`, `READ_API_TOKEN` are long-lived and should be taken
out with special care which is why we have
https://linear.app/unleash/issue/2-1158/add-delete-migration-to-clean-up-no-longer-used-permissions
## Discussion points
If a role has this permission assigned, it will be displayed but will
not be able to remove it. Because the application code does not rely on
these permissions, this shouldn't be a problem. Later when we remove
them from the DB, the permission will be removed as well from the role
by the migration
## About the changes
This PR enables or disables create API token button based on the
permissions.
**Note:** the button is only displayed if you have READ permissions on
some API token. This is a minor limitation as having CREATE permissions
should also grant READ permissions, but right now this is up to the user
to set up the custom role with the correct permissions
**Note 2:** Project-specific API tokens are also ruled by the
project-specific permission to create API tokens in a project (just
having the root permissions to create a client token or frontend token
does not grant access to create a project-specific API token). The
permissions to access the creation of a project-specific API token then
rely on the root permissions to allow the user to create either a client
token or a frontend token.
---------
Co-authored-by: David Leek <david@getunleash.io>
## About the changes
`getUserRootRoles` should also consider custom root roles
This introduces test cases that unveiled a dependency between stores
(this happens actually at the DB layer having access-service access
tables from two different stores but skipping the store layer).
https://linear.app/unleash/issue/2-1161/a-user-with-custom-root-role-and-permission-to-create-client-api
---------
Co-authored-by: Nuno Góis <github@nunogois.com>
This PR changes the OpenAPI schema for the advanced playground to not
accept empty lists of environments and to not accept environment names
that don't match the env name pattern we use.
The pattern is the same as the one we use for controlling environment
names on creation.
However, there is a (small) chance that these may get out of sync later,
so we could do something to only define this pattern once (and import it
in the enterprise package), but that may be more work than is necessary,
and I'd suggest we do that later.
I've also added a minLength to the string items although it isn't
strictly necessary. It's primarily to give the users better feedback if
the name is empty.
This change adds an "edit" link to all playground strategies when they
are returned from the API, allowing the user to jump directly to the
evaluated strategy edit screen.
This change applies to both old and new strategies, so it should even
work in the old playground.
This does not use this information in the frontend yet.
## Discussion points:
Should "links" be an object or a singular string? I know the
notifications service uses just "link", but using an object will make
it easier to potentially add more actions in the future (such as
"enable"/"disable", maybe?)
Do we need to supply basePath? I noticed that the notifications links
only ever use an empty string for base path, so it might not be
necessary and can potentially be left out.
## Changes
I've implemented the link building in a new view model file. Inspecting
the output after the result is fully ready requires some gnarly
introspection and mapping, but it's tested.
Further, I've done a little bit of work to stop the playground service
using the schema types directly as the schema types now contain extra
information.
This PR also updates the `urlFriendlyString` arbitrary to not produce
strings that contain only periods. This causes issues when parsing URLs
(and is also something we struggle with in the UI).
https://github.com/Unleash/unleash/pull/4019 introduced a bug where API
token filtering was not taking into account ADMIN permissions, which
means the API tokens were not being displayed on the UI.
This PR fixes an issue where trying to use an admin token to import
features via the API resulted in a 500 (due to missing properties).
The solution is to catch when the user is using and admin token in the
controller, throw a 400, and tell them to use personal access tokens or
service accounts.
The PR includes a new test file for this specific use case. We don't
really test these cases many other places, so it seemed the logical
choice.
## What
As part of the move to enable custom-root-roles, our permissions model
was found to not be granular enough to allow service accounts to only be
allowed to create read-only tokens (client, frontend), but not be
allowed to create admin tokens to avoid opening up a path for privilege
escalation.
## How
This PR adds 12 new roles, a CRUD set for each of the three token types
(admin, client, frontend). To access the `/api/admin/api-tokens`
endpoints you will still need the existing permission (CREATE_API_TOKEN,
DELETE_API_TOKEN, READ_API_TOKEN, UPDATE_API_TOKEN). Once this PR has
been merged the token type you're modifying will also be checked, so if
you're trying to create a CLIENT api-token, you will need
`CREATE_API_TOKEN` and `CREATE_CLIENT_API_TOKEN` permissions. If the
user performing the create call does not have these two permissions or
the `ADMIN` permission, the creation will be rejected with a `403 -
FORBIDDEN` status.
### Discussion points
The test suite tests all operations using a token with
operation_CLIENT_API_TOKEN permission and verifies that it fails trying
to do any of the operations against FRONTEND and ADMIN tokens. During
development the operation_FRONTEND_API_TOKEN and
operation_ADMIN_API_TOKEN permission has also been tested in the same
way. I wonder if it's worth it to re-add these tests in order to verify
that the permission checker works for all operations, or if this is
enough. Since we're running them using e2e tests, I've removed them for
now, to avoid hogging too much processing time.
As requested in
[Linear](https://linear.app/unleash/issue/2-1147/unleash-cloud-make-keepalive-configurable)
this PR makes the serverKeepAliveTimeout configurable via the
SERVER_KEEPALIVE_TIMEOUT environment variable. This was already
configurable when starting Unleash programmatically, but it's nice to
have as an env variable as well
https://linear.app/unleash/issue/2-1137/roles-unification-on-the-ui
Root and project roles should be managed in a similar manner, which
means using the same roles route and tab for both.
Additionally, this includes a big revamp to the project roles to align
them more closely with the modern and standardized custom root roles
that were recently developed. They mostly use the same components.
There are still more things we want to improve and unify, but we've left
some of that out of this PR due to PR size concerns.
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
Adds an environment variable for switching off feature telemetry in
version check
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
Implements the Advanced Playground Table
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
<!-- Does it close an issue? Multiple? -->
Closes #
[1-1007](https://linear.app/unleash/issue/1-1007/env-aware-results-table)
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
---------
Signed-off-by: andreas-unleash <andreas@getunleash.ai>
This change adds a little more detail to the client metrics schema. The
description for variants felt a
little unclear.
---------
Co-authored-by: Christopher Kolstad <chriswk@getunleash.ai>
## About the changes
Implements custom root roles, encompassing a lot of different areas of
the project, and slightly refactoring the current roles logic. It
includes quite a clean up.
This feature itself is behind a flag: `customRootRoles`
This feature covers root roles in:
- Users;
- Service Accounts;
- Groups;
Apologies in advance. I may have gotten a bit carried away 🙈
### Roles
We now have a new admin tab called "Roles" where we can see all root
roles and manage custom ones. We are not allowed to edit or remove
*predefined* roles.
This meant slightly pushing away the existing roles to `project-roles`
instead. One idea we want to explore in the future is to unify both
types of roles in the UI instead of having 2 separate tabs. This
includes modernizing project roles to fit more into our current design
and decisions.
Hovering the permissions cell expands detailed information about the
role:
### Create and edit role
Here's how the role form looks like (create / edit):
Here I categorized permissions so it's easier to visualize and manage
from a UX perspective.
I'm using the same endpoint as before. I tried to unify the logic and
get rid of the `projectRole` specific hooks. What distinguishes custom
root roles from custom project roles is the extra `root-custom` type we
see on the payload. By default we assume `custom` (custom project role)
instead, which should help in terms of backwards compatibility.
### Delete role
When we delete a custom role we try to help the end user make an
informed decision by listing all the entities which currently use this
custom root role:
~~As mentioned in the screenshot, when deleting a custom role, we demote
all entities associated with it to the predefined `Viewer` role.~~
**EDIT**: Apparently we currently block this from the API
(access-service deleteRole) with a message:
What should the correct behavior be?
### Role selector
I added a new easy-to-use role selector component that is present in:
- Users
- Service Accounts
- Groups
### Role description
I also added a new role description component that you can see below the
dropdown in the selector component, but it's also used to better
describe each role in the respective tables:
I'm not listing all the permissions of predefined roles. Those simply
show the description in the tooltip:
### Role badge
Groups is a bit different, since it uses a list of cards, so I added yet
another component - Role badge:
I'm using this same component on the profile tab:
## Discussion points
- Are we being defensive enough with the use of the flag? Should we
cover more?
- Are we breaking backwards compatibility in any way?
- What should we do when removing a role? Block or demote?
- Maybe some existing permission-related issues will surface with this
change: Are we being specific enough with our permissions? A lot of
places are simply checking for `ADMIN`;
- We may want to get rid of the API roles coupling we have with the
users and SAs and instead use the new hooks (e.g. `useRoles`)
explicitly;
- We should update the docs;
- Maybe we could allow the user to add a custom role directly from the
role selector component;
---------
Co-authored-by: Gastón Fournier <gaston@getunleash.io>
## About the changes
Unleash version should be the one best representing its runtime. In this
regard, enterpriseVersion trumps unleash-server version.
This version is the one that's sent as part of our metrics.
<!-- Thanks for creating a PR! To make it easier for reviewers and
everyone else to understand what your changes relate to, please add some
relevant content to the headings below. Feel free to ignore or delete
sections that you don't think are relevant. Thank you! ❤️ -->
## About the changes
<!-- Describe the changes introduced. What are they and why are they
being introduced? Feel free to also add screenshots or steps to view the
changes if they're visual. -->
Adds feature usage info and custom strategy counters to the version
check object.
<!-- Does it close an issue? Multiple? -->
Closes #
<!-- (For internal contributors): Does it relate to an issue on public
roadmap? -->
<!--
Relates to [roadmap](https://github.com/orgs/Unleash/projects/10) item:
#
-->
### Important files
<!-- PRs can contain a lot of changes, but not all changes are equally
important. Where should a reviewer start looking to get an overview of
the changes? Are any files particularly important? -->
## Discussion points
<!-- Anything about the PR you'd like to discuss before it gets merged?
Got any questions or doubts? -->
## About the changes
When a feature is not found in a project we should fail with a NotFound
error. If the feature belongs to a different project, it should not be a
permission issue, because the user might not be aware (lack of
permissions/visibility) of that other project, so even in this case the
error should be NotFound (this also works if we ever allow the same
feature name in different projects)
Fixes#3726
---------
Co-authored-by: Thomas Heartman <thomas@getunleash.ai>
## About the changes
Edit application under
https://app.unleash-hosted.com/demo/applications/test-app is currently
not working as the appName is expected to come in the request body, but
it's actually part of the url. We have two options here:
1. We change the UI to adapt to the expectations of the request by
adding appName to the request body (and eventually removing appName from
the URL, which would be a breaking change)
2. We remove the restriction of only sending the appName in the body and
take the one that comes in the URL. We have a validation that verifies
that at least one of the two sets the appName
([here](e376088668/src/lib/services/client-metrics/instance-service.ts (L208))
we validate using [this
schema](e376088668/src/lib/services/client-metrics/schema.ts (L55-L70)))
In terms of REST API, we can assume that the appName will be present in
the resource `/api/admin/metrics/applications` (an endpoint we don't
have), but when we're updating an application we should refer to that
application by its URL: `/api/admin/metrics/applications/<appName>` and
the presence of an appName in the body might indicate that we're trying
to change the name of the application (something we currently not
support)
Based on the above, I believe going with the second option is best, as
it adheres to REST principles and does not require a breaking change.
Despite that, we only support updating applications as the creation is
done from metrics ingestion
Fixes: #3580
This PR aims to handle the increased log alarm volume seen by the SREs.
It appears that we get a large number of alarms because a client
disconnects early from the front-end API. These errors are then
converted into 500s because of missing handling. These errors appear to
be caused by the `http-errors` library in a dependency.
We also introduced a log line whenever we see errors now a while back,
and I don't think we need this logging (I was also the one who
introduced it).
The changes in this PR are specifically:
- When converting from arbitrary errors, give `BadRequestError` a 400
status code, not a 500.
- Add a dependency on `http-errors` (which is already a transitive
dependency because of the body parser) and use that to check whether an
error is an http-error.
- If an error is an http error, then propagate it to the user with the
status code and message.
- Remove warning logs when an error occurs. This was introduced to make
it easier to correlate an API error and the logs, but the system hasn't
been set up for that (yet?), so it's just noise now.
- When logging errors as errors, only do that if their status code would
be 500.
This change makes the logs that happen when we encounter an error a
little bit clearer. It logs the error message before the error ID and
also logs the full serialized message just in case.
This PR improves our handling of internal Joi errors, to make them more
sensible to the end user. It does that by providing a better description
of the errors and by telling the user what they value they provided was.
Previous conversion:
```json
{
"id": "705a8dc0-1198-4894-9015-f1e5b9992b48",
"name": "BadDataError",
"message": "\"value\" must contain at least 1 items",
"details": [
{
"message": "\"value\" must contain at least 1 items",
"description": "\"value\" must contain at least 1 items"
}
]
}
```
New conversion:
```json
{
"id": "87fb4715-cbdd-48bb-b4d7-d354e7d97380",
"name": "BadDataError",
"message": "Request validation failed: your request body contains invalid data. Refer to the `details` list for more information.",
"details": [
{
"description": "\"value\" must contain at least 1 items. You provided [].",
"message": "\"value\" must contain at least 1 items. You provided []."
}
]
}
```
## Restructuring
This PR moves some code out of `unleash-error.ts` and into a new file.
The purpose of this is twofold:
1. avoid circular dependencies (we need imports from both UnleashError
and BadDataError)
2. carve out a clearer separation of concerns, keeping `unleash-error` a
little more focused.
This change lowers the log level from warning to debug for when we see
unexpected error types.
Right now this triggers for Joi errors, which we still rely on pretty
heavily. Lowering this should clear up logs for most users.
If apiTokens are enabled breaks middleware chain with a 401 if no token
is found for requests to client and frontend apis. Previously the
middleware allowed the chain to process.
Removes the regex search for multiple slashes, and instead configures
the apiTokenMiddleware to reject unauthorized requests.
