This PR aims to improve performances as we are making A LOT of queries
after having added events emission to the ORM layer. While investigating
issues, I've come across multiple problems I will describe below
## Add index on workspace.activationStatus as we are querying it a lot
As per title
## Add logs on core datasource destroy
It seems that we have postgres connection pool destruction in
production. I cannot reproduce locally but I suspect the Query Timeouts
to be the root cause. I'm fixing most of the Query Timeouts cause in
this PR but I'm adding the logs so we have more information in case in
keeps happening in production.
## GraphQL query runner createMany
It was using a for loop on each record. This is as issue has we will
emit an event separately for each record => we should always try to
batch events.
Replacing by a save on all records. Note that this is not perfect as we
should avoid using save (bad performances), and use insert + updateMany
instead. As I have follow up discussions regarding permissions, I
haven't replaced it by insert and updateMany yet. Using save instead of
a for loop makes the problem less worrying.
## Introduce updateMany in ORM @Weiko @ijreilly FYI
.save(manyRecords) is bad as it's querying the data for no good reason
(and doing a select for each record...).
We already have .insert(), i'm introducing .updateMany()
I think our ORM layer should be simplified a lot but I'm not starting
the refacto yet
## Fixing ORM @Weiko @ijreilly FYI
- Fixing events emission in delete function
- Fixing events emission in insert function
- make sure everything is batched
## Events performance @Weiko @ijreilly FYI
Do not emit timelineActivity db events as this does not seem useful and
is quite heavy
## Messaging and Calendar performance @bosiraphael FYI
Rework many functions to make sure they are batched. This does not touch
the driver layer and I have heavily tested it. (multiple account, with
multiple channels, common thread, common messages, etc...)
## Workflow Trigger relation Fetch performance @martmull FYI
Improved performance by batching
Fixes https://github.com/twentyhq/core-team-issues/issues/1262
In this PR we add the update permission check layer by
- for the graphql api: extracting columns to update from the
expressionMap
- for rest api: .save() is used so we need to add the permission layer
to .save directly. We also take advantage of this PR to filter out
non-readable fields from save response (other save returns the whole
entity) - this was planned in
https://github.com/twentyhq/core-team-issues/issues/1216
The current solution does not work with rest api depth 2 queries, but
this seem to already not work on main (for timeout reasons though, so
different). I offer to create a ticket to fix it altogether later.
Context :
Large PR with 600+ test files. Enable connect and disconnect logic in
createMany (upsert true) / updateOne / updateMany resolvers
- Add disconnect logic
- Gather disconnect and connect logic -> called relation nested queries
- Move logic to query builder (insert and update one) with a preparation
step in .set/.values and an execution step in .execute
- Add integration tests
Test :
- Test API call on updateMany, updateOne, createMany (upsert:true) with
connect/disconnect
This PR adapts teh deleteOneField to make sure morph relations can be
deleted, either
- from a relation type,
- or from a morph relation type (which is the most obvious case).
This PR covers
- the deletion of fieldMetadata
- and the migrationof workspace schemas, by using the already existing
definition of relation migrations
This PR implements a new test suite: "Delete Object metadata with morph
relation" and completes the other test suites for FieldMetadata and
ObjectMetadata creation/deletion.
Last, we added a nitpick from @paul I forgot on a previous PR on
process-nested-realtion-v2
Fixes https://github.com/twentyhq/core-team-issues/issues/1197
- Renamed `WorkflowActionAdapter` to `ToolExecutorWorkflowAction`
- Renamed `settingPermission` table to `permissionFlag` and `setting`
column to `flag`
- Decoupled the send email logic from workflows to tools
- Add new `Tools Permission` section in FE
---------
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
Co-authored-by: Félix Malfait <felix@twenty.com>
In this PR, behind a feature flag, we add a permission layer check based
on the read permission.
It is done by computing a map of an object's fields, where keys are the
column names and values the fieldMetadata id, making them comparable to
the restricted fields ids list stored in the permission cache.
For mutations (create, update, delete, destroy), we need to check the
read permission on the returned field, as they may differ from the
updated field. The write field permission will be tackled in a different
PR.
## Problem
After migrating webhooks and API keys from workspace to core level, REST
API endpoints were still creating entities in workspace schema
(`workspace_*`) instead of core schema, causing webhooks to not fire.
## Solution
- Added dedicated REST controllers for webhooks (`/rest/webhooks`) and
API keys (`/rest/apiKeys`)
- Updated dynamic controller to block workspace-gated entities from
being processed
- Fixed OpenAPI documentation to exclude these endpoints from playground
- Ensured return formats match GraphQL resolvers exactly
## Testing
✅ All endpoints tested with provided auth token - webhooks and API keys
now correctly stored in `core` schema
# Introduction
Following https://github.com/twentyhq/twenty/pull/13264, this PR
introduces several `fieldMetadataEntity` typing enhancement suggestions.
Mainly any nullable field metadata entity properties are now either
nullable or defined.
Or never if field is dynamically required or not depending on the field
metadata type
This enhance DevX
## Standards
- field enum ( `MULTI_SELECT`, `SELECT`, `RATING` ) will never have
`options` set to `NULL` in db
- field `RELATION` or `MORH_RELATION` won't ever have its relation
fields set to `NULL` in db
- field of any type `settings`, even if possibly defined, can still be
`NULL` in db
- field of any type `defaultValue`, even if possibly defined, can still
be `NULL` in db
It coud be interesting to guard these standards by adding dedicated pg
constraints on each field
## TypesScript type tests
added coverage for each `settings`, `defaultValue`, and `options`
depending on the current `fieldMetadata`
Honestly I don' know if this typescript assertions test file is not
overkill, but regarding metadata staticness it might be very interesting
to have this guard
## Possible improvements
- We could type as `unknown` instead of "all" on `FieldMetadataType`
inferrance
- We still need to deprecate remaining duplicated entities such as
`Index/Field/MetadataInterface` etc not a huge refactor neither urgent
# Introduction
Following https://github.com/twentyhq/twenty/pull/13310
> After this PR merge will create a new one removing the type and
replacing it to ObjectMetadataEntity.
This is it !
Short fix for the morph case.
Was missing the logic for the joinColumnName: we need to add the
"ObjcectMetadataName" on top the the "FieldMetadataName" for morph
relation like it was designed on fieldmetadata service.
On the typescirpt side: the "as FieldMetadataRelationSettings" is not
ideal but I inherit from the a as FieldMetadataInterface type that does
not know the as FieldMetadataType from its parent. If you have a better
simple idea, let me know.
Note : I think we could refactor a bit this part later on.
# Introduction
Following `FieldMetadataInterface` deprecation in
https://github.com/twentyhq/twenty/pull/13264
As for the previous PR will rename and remove all the file in a
secondary PR to avoid conflicts and over loading this one
## Improvements
Removed optional properties from the `objectMetadataEntity` model and
added utils to retrieve test data
## Notes
By touching to `ObjectMetadataDTO` I would have expected a twenty-front
codegenerated types mutation, but it does not seem to be granular enough
to null/undefined coercion
# Introduction
Following https://github.com/twentyhq/twenty/pull/13264
> After this PR merge will create a new one removing the type and
replacing it to FieldMetadataEntity.
This is it !
# Introduction
From the moment replaced the FieldMetadataInterface definition to:
```ts
import { FieldMetadataType } from 'twenty-shared/types';
import { FieldMetadataEntity } from 'src/engine/metadata-modules/field-metadata/field-metadata.entity';
export type FieldMetadataInterface<
T extends FieldMetadataType = FieldMetadataType,
> = FieldMetadataEntity<T>;
```
After this PR merge will create a new one removing the type and
replacing it to `FieldMetadataEntity`.
Did not renamed it here to avoid conflicts on naming + type issues fixs
within the same PR
## Field metadata entity RELATION or MORPH
Relations fields cannot be null for those field metadata entity instance
anymore, but are never for the others see
`packages/twenty-server/src/engine/metadata-modules/field-metadata/types/field-metadata-entity-test.type.ts`
( introduced TypeScript tests )
## Concerns
- TS_VECTOR is the most at risk with the `generatedType` and
`asExpression` removal from interface
## What's next
- `FielMetadataInterface` removal and rename ( see introduction )
- Depcrecating `ObjectMetadataInterface`
- Refactor `FieldMetadataEntity` optional fiels to be nullable only
- TO DIG `never` occurences on settings, defaultValue etc
- Some interfaces will be replaced by the `FlatFieldMetadata` when
deprecating the current sync and comparators tools
Fixes: https://github.com/twentyhq/twenty/issues/13110
I'm deprecating note.body and task.body to remove confusion between body
and bodyV2
What will be left but should be done later to avoid breaking changes:
- re-add a body field in the graphql API only that points to the same
bodyV2 field in SQL (need to be handled in fields and filter for note
and task)
- (wait some time)
- remove bodyV2 field
Issue introduced by [Restrict queried columns to graphql-requested
fields](https://github.com/twentyhq/twenty/pull/13246)
In this query
```ts
{
...
name
messageId
message {
id
}
}
```
`messageId` was being filtered out from the selected fields as we failed
to link it to an existing field, thus null was always returned.
`message { id }` worked because we already handled connections
correctly.
## Context
Add an eventEmitter instance to twenty datasources so we can emit DB
events.
Add input and output formatting to twenty orm (formatData, formatResult)
Those 2 elements simplified existing logic when we interact with the
ORM, input will be formatted by the ORM so we can directly use
field-like structure instead of column-like. The output will be
formatted, for builder queries it will be in `result.generatedMaps`
where `result.raw` preserves the previous column-like structure.
Important change: We now have an authContext that we can pass when we
get a repository, this will be used for the different events emitted in
the ORM. We also removed the caching for repositories as it was not
scaling well and not necessary imho
Note: An upcoming PR should handle the onDelete: cascade behavior where
we send DESTROY events in cascade when there is an onDelete: CASCADE on
the FK.
---------
Co-authored-by: Charles Bochet <charles@twenty.com>
Fixes
https://github.com/twentyhq/core-team-issues/issues/255?issue=twentyhq%7Ccore-team-issues%7C1214.
Until then, in the endpoints of our dynamic schema, we were querying all
columns and then formatting the result by removing the non-requested
fields (fields not mentioned in the graphql Query) from the result.
This is not compatible with field-level permissions that we are about to
introduce because users would see their request denied if they have
restricted rights on any of the fields of the objects they are querying,
even if they did not query it in the first place.
To prepare for this change, we are restricting the list of queried
columns to those made necessary by the graphql query.
I only made the changes in the dynamic schema for now. We will
potentially need to do updates to other part of the app that use
createQueryBuilder directly or not (for instance, when calling
repository methods such as .findOne()), but they mostly regard system
objects that are not subject to permissions or are executed by entities
that bypass permission such as jobs creating People and Companies from
their email sync.
No changes have been brought to existingRecords related logic in the
dynamic schema because @Weiko is currently working on it, so I may need
to adapt the new logic after he is done.
No feature flag have been added so far as this should not change
anything at the moment.
Context:
Users are complaining to see their workspace in a language they don't
know. This behavior is transient, happens on data model update and
disappear on refresh
I've check the cache for users that got the issue and did not spot any
weird language
==> I think we somehow fallback the the request header locale. I feel we
should always use the userWorkspace.locale, request locale should not be
used in BE in my opinion except for unauthenticated endpoints. I'm also
adding logs to understand the locale issue
In this PR:
rename user.workspaces into user.userWorkspaces which is more correct
improve / simplify LOCALES typing
In the BE we throw custom errors with precise error codes (e.g.
"LABEL_ALREADY_EXISTS") before catching them in filters and rethrowing
BaseGraphQLErrors (standard errors such as NotFoundError, UserInputError
etc.).
In the FE we were grouping sentries based on the error codes but we were
actually grouping by very broad codes such as "NOT_FOUND" or
"BAD_USER_INPUT", extracted from the BaseGraphQLErrors.
To fix that, we update the BaseGraphQLError constructor api to allow to
pass on the CustomError directly and retrieve from it the original code
and store it in existing property `subCode` that we will use in the FE
to send errors to sentry.
This new api also eases usage of `userFriendlyMessage` that is passed on
to the api response and therefore to the FE when CustomError is passed
on directly to the BaseGraphQLError constructor.
Fixes https://github.com/twentyhq/twenty/issues/12867
Issue:
when you have a variable `toto` which is: `Record<string, MyType>` and
you do toto['xxx'], this will be typed as `MyType` instead of `MyType |
undefined`
Solutions:
- activate `noUncheckedIndexedAccess` check in tsconfig, this is the
preferred solution but will take time to get there (this raises 600+
errors)
- use a Map: cf https://github.com/twentyhq/twenty/pull/13125/files
- set the type to Partial<Record<string, MyType>>. Drawback is that when
you do Object.values(toto), you'll get `Array<MyType | undefined>`.
Hence why we have to filter these behind
<img width="1512" alt="image"
src="https://github.com/user-attachments/assets/d0a0bfed-c441-4e53-84c2-2da98ccbcf50"
/>
Large PR, sorry for that. Don't hesitate to reach me to have full
context (env. 500lines for integration and unit tests)
- Add connect logic in Workspace Entity Manager
- Update QueryDeepPartialEntity type to enable dev to use connect
- Add integration test on createOne / createMany
- Add unit test to cover main utils
- Remove feature flag on connect
closes https://github.com/twentyhq/core-team-issues/issues/1148
closes https://github.com/twentyhq/core-team-issues/issues/1147
In this PR, I'm fixing a bug introduced in recent performance work on
the cache.
Bug context: https://github.com/twentyhq/twenty/issues/12865
Related PR opened by a contributor:
https://github.com/twentyhq/twenty/pull/13003
## Root cause
We cache all objectMetadataItems at graphql level : see
`useCachedMetadata` hook:
- instead of going through the regular resolvers, we direlcty load data
from the cache. However this data must be localized regarding labels and
descriptions
In a precedent refactoring, we introduced the notion of locale in the
cache key. However, the user locale was not properly taken into account
as we did not have the information in this hook.
## Fix
1. **Introduce locale in userWorkspace entity**. The locale is stored on
workspaceMember in each postgres workspaceSchema (workspace_xxx) which
is the alter ego of userWorkspace in postgres core schema. Note that we
can't store it in user as a user can be part of multiple workspaces (the
locale already there must be seen as a default for this user), and we
cannot rely on workspaceMember as we would need to query the
workspaceSchema in the authentication layer which we want to avoid for
performance reasons.
2. During request hydration from token (containing the userWorkspaceId),
we fetch the userWorkspace and store it in the Request (this impact both
AuthContext and Request interface)
3. Leverage userWorkspace.locale in the useCachedMetadata hook
## Additional notes
There is no need to change the way we store and retrieve the
object-metadata-maps object itself which is different from the graphql
layer cache. object-metadadata-maps are not localized
This PR introduces a significant enhancement to the role-based
permission system by extending it to support AI agents, enabling them to
perform database operations based on assigned permissions.
## Key Changes
### 1. Database Schema Migration
- **Table Rename**: `userWorkspaceRole` → `roleTargets` to better
reflect its expanded purpose
- **New Column**: Added `agentId` (UUID, nullable) to support AI agent
role assignments
- **Constraint Updates**:
- Made `userWorkspaceId` nullable to accommodate agent-only role
assignments
- Added check constraint `CHK_role_targets_either_agent_or_user`
ensuring either `agentId` OR `userWorkspaceId` is set (not both)
### 2. Entity & Service Layer Updates
- **RoleTargetsEntity**: Updated with new `agentId` field and constraint
validation
- **AgentRoleService**: New service for managing agent role assignments
with validation
- **AgentService**: Enhanced to include role information when retrieving
agents
- **RoleResolver**: Added GraphQL mutations for `assignRoleToAgent` and
`removeRoleFromAgent`
### 3. AI Agent CRUD Operations
- **Permission-Based Tool Generation**: AI agents now receive database
tools based on their assigned role permissions
- **Dynamic Tool Creation**: The `AgentToolService` generates CRUD tools
(`create_*`, `find_*`, `update_*`, `soft_delete_*`, `destroy_*`) for
each object based on role permissions
- **Granular Permissions**: Supports both global role permissions
(`canReadAllObjectRecords`) and object-specific permissions
(`canReadObjectRecords`)
### 4. Frontend Integration
- **Role Assignment UI**: Added hooks and components for
assigning/removing roles from agents
## Demo
https://github.com/user-attachments/assets/41732267-742e-416c-b423-b687c2614c82
---------
Co-authored-by: Antoine Moreaux <moreaux.antoine@gmail.com>
Co-authored-by: Lucas Bordeau <bordeau.lucas@gmail.com>
Co-authored-by: Charles Bochet <charles@twenty.com>
Co-authored-by: Guillim <guillim@users.noreply.github.com>
Co-authored-by: Charles Bochet <charlesBochet@users.noreply.github.com>
Co-authored-by: Weiko <corentin@twenty.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: github-actions <github-actions@twenty.com>
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
Co-authored-by: Marie <51697796+ijreilly@users.noreply.github.com>
Co-authored-by: martmull <martmull@hotmail.fr>
Co-authored-by: Thomas Trompette <thomas.trompette@sfr.fr>
Co-authored-by: Etienne <45695613+etiennejouan@users.noreply.github.com>
Co-authored-by: Baptiste Devessier <baptiste@devessier.fr>
Co-authored-by: nitin <142569587+ehconitin@users.noreply.github.com>
Co-authored-by: Paul Rastoin <45004772+prastoin@users.noreply.github.com>
Co-authored-by: prastoin <paul@twenty.com>
Co-authored-by: Vicky Wang <157669812+vickywxng@users.noreply.github.com>
Co-authored-by: Vicky Wang <vw92@cornell.edu>
Co-authored-by: Raphaël Bosi <71827178+bosiraphael@users.noreply.github.com>
Context :
- IndexFieldMetadata was no longer available on 'objects' gql query
([since this PR](https://github.com/twentyhq/twenty/pull/12785)). Then,
unicity checks on import do not work anymore.
Fix :
- Add a dataloader logic in indexFieldMetadata
- Add extra check in unicity hook on import
## Context
Added to the existing useGraphQLErrorHandlerHook yoga hook to increment
metrics after all query executions based on their error codes. I
originally wanted to create a new useMetrics hook but most of the error
handling was done in useGraphQLErrorHandlerHook so we decided to keep it
there for now.
<img width="1310" alt="Screenshot 2025-06-24 at 15 58 26"
src="https://github.com/user-attachments/assets/498d3754-851a-4051-a5c2-23ac8253aa6a"
/>
In this PR:
- add query hashKey to ObjectMetadataItems query graphql cache to avoid
caching outdated queries
- improve performance by removing ResolveField at FieldLevel and adding
this at resolver level
In this PR:
## Improve recompute metadata cache performance. We are aiming for
~100ms
Deleting relationMetadata table and FKs pointing on it
Fetching indexMetadata and indexFieldMetadata in a separate query as
typeorm is suboptimizing
## Remove caching lock
As recomputing the metadata cache is lighter, we try to stop preventing
multiple concurrent computations. This also simplifies interfaces
## Introduce self recovery mecanisms to recompute cache automatically if
corrupted
Aka getFreshObjectMetadataMaps
## custom object resolver performance improvement: 1sec to 200ms
Double check queries and indexes used while creating a custom object
Remove the queries to db to use the cached objectMetadataMap
## reduce objectMetadataMaps to 500kb
<img width="222" alt="image"
src="https://github.com/user-attachments/assets/2370dc80-49b6-4b63-8d5e-30c5ebdaa062"
/>
We used to stored 3 fieldMetadataMaps (byId, byName, byJoinColumnName).
While this is great for devXP, this is not great for performances.
Using the same mecanisme as for objectMetadataMap: we only keep byIdMap
and introduce two otherMaps to idByName, idByJoinColumnName to make the
bridge
## Add dataloader on IndexMetadata (aka indexMetadataList in the API)
## Improve field resolver performances too
## Deprecate ClientConfig
