**AI Chat - Tool Executions (counters, tagged with model)**
ai-chat/tool-execution-succeeded: number of tool calls invoked by the AI
that completed without error
ai-chat/tool-execution-failed: number of tool calls invoked by the AI
that threw an error
**AI Chat - Token Usage (counters, tagged with model)**
ai-chat/input-tokens: total input tokens sent to the model across all
turns
ai-chat/output-tokens: total output tokens generated by the model
ai-chat/cache-read-tokens: input tokens served from the model's prompt
cache (cheaper)
ai-chat/cache-write-tokens: input tokens written into the prompt cache
for future reuse
**AI Chat - Latency (histograms in ms, tagged with model)**
ai-chat/turn-latency-ms: total duration of a full chat turn (from stream
start to stream end)
ai-chat/step-latency-ms: duration of a single reasoning/tool-call step
within a turn
ai-chat/ttft-ms: time-to-first-token, i.e. how long until the model
starts streaming output
**MCP - Tool Executions (counters)**
mcp/tool-execution-succeeded: number of MCP tool calls that completed
successfully
mcp/tool-execution-failed: number of MCP tool calls that threw an error
## Summary
- Add `WORKSPACE_SCHEMA_DDL_LOCKED` env-only boolean config variable
that blocks all workspace schema DDL changes when set to `true`. This is
intended for hot upgrades where logical replication cannot handle DDL
changes. Enforced at two chokepoints:
- `WorkspaceMigrationRunnerService.run` — blocks all metadata-driven DDL
(object/field/index CRUD, app sync/uninstall, standard app sync, upgrade
commands)
- `WorkspaceDataSourceService.createWorkspaceDBSchema` /
`deleteWorkspaceDBSchema` — blocks workspace creation (sign-up) and hard
deletion. Uses a dedicated `WorkspaceDataSourceException` (not
ForbiddenException)
- Add maintenance mode feature with Admin Panel UI and user-facing
banner:
- **Backend**: `MaintenanceModeService` stores maintenance window
(startAt, endAt, optional link) in `core.keyValuePair` as
`CONFIG_VARIABLE`. Validates endAt > startAt. Uses `GraphQLISODateTime`
scalar for date fields. Exposed via `clientConfig` REST endpoint and
admin GraphQL mutations (`setMaintenanceMode`, `clearMaintenanceMode`)
- **Admin Panel**: New "Maintenance Mode" section in Health tab with UTC
datetime pickers and activate/deactivate controls
- **Banner**: `InformationBannerMaintenance` displayed at the top of
`DefaultLayout` for all users, using Temporal API for timezone-aware
formatting with an optional "Learn more" link
These two features are **independent** — the DDL lock is controlled via
env var for operational use, while maintenance mode is a UI notification
mechanism controlled from the admin panel.
Fixes https://github.com/twentyhq/twenty/issues/13103
Newly created workspaces start with metadataVersion = 0. The truthy
check this.currentWorkspace?.metadataVersion && {…} treated 0 as falsy,
so the X-Schema-Version header was never sent. Without this header, the
backend's schema mismatch detection is bypassed and raw validation
errors leak to Sentry. Replaced with isDefined() check.
Also replaced bare negation checks with isDefined() in the backend
validation hook for consistency.
Added a new IS_GRAPHQL_QUERY_TIMING_ENABLED feature flag that, when
activated per workspace, logs the execution time and operation name of
every GraphQL query across both the standard Yoga pipeline and the
direct execution fast path. The timing context propagates via
AsyncLocalStorage to also instrument buildColumnsToSelect and
formatResult — giving a breakdown of column selection and result
transformation costs without any caller changes.
# Introduction
Previously the auth jwt stragegy would lod the whole user entity in the
auth user context
On an exception it would completely get logged on the pods
## Security layer
- 0/ Updating the type system ( devxp only though )
- 1/ The jwt auth stragegy only load a specific sub set of the user
entity
- 2/ Sanitizing at the exception log level directly in case of a user
context
- 3/ Sanitizing at the console driver
The last two sanitization could sound a bit redundant though they're
still good fallback to keep in case new path occurs in the cb
This PR handles SSE event stream edge cases.
- When a pod restarts, the front clients have to reconnect to SSE
- When the dev server restarts or is hot reloaded, the front client has
to reconnect to SSE
- When redis server restarts or the redis key is cleared for any reason,
the server has to recreate the event stream in redis, this can happen
when navigating for example.
- Log in / log out flow
With this PR we avoid error messages in the front end due to TTL or pod
crash, we implement a resilient way of reconnecting silently.
To avoid DDoSing our servers if pods crash or a full restart of the
cluster is made, we evenly space retry attempts to reconnect from all
the clients, to avoid n clients reconnection at the same time, we use a
random wait time between 0 and a constant max wait time (set to 2 mins
for now). This is the cheapest and most effective solution, clients who
want to force reconnect have to refresh or navigate to another page.
Fixes https://github.com/twentyhq/core-team-issues/issues/2045
# Introduction
In this PR we catch all the runner errors coming from a single workspace
migration action execution.
**1. `WorkspaceMigrationActionExecutionException`** (low-level,
action-specific)
- Thrown from action handlers, utils, and helper functions
- Contains specific error codes: `FIELD_METADATA_NOT_FOUND`,
`OBJECT_METADATA_NOT_FOUND`, `ENUM_OPERATION_FAILED`, `NOT_SUPPORTED`,
etc.
- Simple structure: `message`, `code`, `userFriendlyMessage`
- No action context - just describes what went wrong
**2. `WorkspaceMigrationRunnerException`** (high-level, runner-scoped)
- Only two codes: `INTERNAL_SERVER_ERROR` and `EXECUTION_FAILED`
- `EXECUTION_FAILED` **requires** `action` + `errors` (contains the
action context)
- `INTERNAL_SERVER_ERROR` **requires** `message` (no action context)
## Refactor
- Removed the `relatedFlatEntityMapsKeys` from the `WorkspaceMigration`
type as they're directly inferred from passed actions
- Swallowing actions rollbacks errors in order to iterate over all of
them
## Testing
Created a very straigthforward install application from workspace
migration endpoint in order to start testing the introduced
`WorkspaceMigrationActionExecutionException`
Introduced a feature flag that stop the access to the endpoint if not
enabled
Whole taken direction are totally subjective and highly prone to
mutations ( endpoint location, naming and input schema see
`ts-expect-error` comment )
cc @martmull
## Response error
```ts
{
"eventId": "evt_a1b2c3d4-5678-90ab-cdef-1234567890ab",
"extensions": {
"action": {
"metadataName": "fieldMetadata",
"type": "delete",
"universalIdentifier": "20202020-6110-4547-9fd0-2525257a2c3f"
},
"code": "APPLICATION_INSTALLATION_FAILED",
"errors": {
"metadata": {
"code": "ENTITY_NOT_FOUND",
"message": "Could not find flat entity with universal identifier 20202020-6110-4547-9fd0-2525257a2c3f"
},
"workspaceSchema": {
"code": "ENTITY_NOT_FOUND",
"message": "Could not find flat entity in maps"
}
},
"exceptionEventId": "exc_f9e8d7c6-5432-10ba-fedc-ba0987654321",
"userFriendlyMessage": "Migration execution failed."
},
"message": "Migration action 'delete' for 'fieldMetadata' failed",
"name": "GraphQLError"
}
```
## Context
After flushing the Redis cache (e.g., via `cache:flush` command), users
get stuck in an infinite loop showing "Your workspace has been updated
with a new data model. Please refresh the page." - refreshing the page
doesn't help.
## Root Cause
When the cache is flushed, `getMetadataVersion()` returns `undefined`.
The version comparison in the error handler then becomes:
```typescript
requestMetadataVersion !== `${currentMetadataVersion}`
// Evaluates to: "5" !== "undefined" → always true
```
This triggers the schema mismatch error on every request, even after
page refresh.
## History
| Date | Commit | What Happened |
|------|--------|---------------|
| Aug 2024 | #6691 (`17a1760afd`) | Introduced the
`${currentMetadataVersion}` template literal that converts `undefined`
to the string `"undefined"` |
| Dec 2025 | #16536 (`0035fc1145`) | Removed the safety check that would
throw an early error when cache is empty, allowing `undefined` to
propagate |
## Fix
Add `isDefined(currentMetadataVersion)` check before comparing versions.
If the server doesn't have a cached version, we shouldn't treat it as a
mismatch - the schema factory already handles populating the cache when
it actually needs the version.
```typescript
if (
requestMetadataVersion &&
isDefined(currentMetadataVersion) && // ← Added this check
requestMetadataVersion !== `${currentMetadataVersion}`
) {
```
## Testing
1. Flush the cache: `npx nx run twenty-server:command cache:flush`
2. Refresh the page
3. Verify no infinite loop occurs and app loads normally
Validations :
- relations count (in common api)
- oneToMany relation nested count (in common)
- requested fields count (in gql)
- root resolver count (in gql)
- root resolver duplicates (in gql)
- specific complexity for metadata / nesting count (in gql)
This PR fixes a crash of the app when a Snackbar tries to render an
object.
<img width="3002" height="754" alt="image"
src="https://github.com/user-attachments/assets/988f97eb-5c8a-44f8-ac8e-e2953b872bac"
/>
What happened : the backend sent a MessageDescriptor in
`extensions.userFriendlyMessage` while it should have sent a translated
string.
But it is a problem that the Snackbar can end up rendering objects and
crashing the whole app because it is a the highest level in the tree.
So this PR hardens many points to avoid future bugs :
- Sanitize what is rendered by the Snackbar to avoid any crash
- Translate any MessageDescriptor object that could end up in the
frontend
- Fix the places in the backend where a MessageDescriptor wasn't
translated and sent directly to the frontend in
`use-graphql-error-handler.hook.ts`
Fixes https://github.com/twentyhq/twenty/issues/15685
Fixes https://github.com/twentyhq/core-team-issues/issues/1869
I had an issue with invalid UUIDs, and I think it would be easier to
find the offending one if the UUID were included in the error message.
This way a database dump can be easily searched.
---------
Co-authored-by: prastoin <paul@twenty.com>
Following https://github.com/twentyhq/twenty/pull/14869
Dynamically clearing the cache, as one entry was forgotten
It seems like a permission test deletes a role that's expected to be
existing in following tests, as now cache is getting invalidated tests
are failing
Seems to be related to
https://discord.com/channels/1130383047699738754/1423768505911869460
## Singleton local cache key collision
When set for the first time caches local keys looks like
`workspaceId:undefined", singleton is shared between several cache
instances
If the given key is undefined it will read on other entity cache entry.
## Improvements
- Add logs to all gql operations and rest calls to help debug CPU issues
on the backend. These are temporary and should be removed
- Remove nested relations from workflowVersions load (used to add manual
triggers in the side bar). On some workspaces this call result in a
response of 4MB which is heavy on CPU
- Investigated Redis Usage ==> made a few improvements, we are should
still migrate to the new cache service once available
- investigated db calls in messaging / calendar fetch list + workflow
enqueue run cron jobs. Everything seems to be properly batched
We decided to send 4xx errors to sentry from the FE, except for
ForbiddenErrors.
In some cases, we also need not to send to sentry some 4xx errors, like
user input error, because the error stemming from the user input is
unpredictable (ex: duplicate entry for a record with unicity
constraints) and thus acceptable.
Let's add a isExpected extension on UserInputErrors to introduce a
specific behaviour for them (do not send to sentry).
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.
Currently, when a server query or mutation from the front-end fails, the
error message defined server-side is displayed in a snackbar in the
front-end.
These error messages usually contain technical details that don't belong
to the user interface, such as "ObjectMetadataCollection not found" or
"invalid ENUM value for ...".
**BE**
In addition to the original error message that is still needed (for the
request response, debugging, sentry monitoring etc.), we add a
`displayedErrorMessage` that will be used in the snackbars. It's only
relevant to add it for the messages that will reach the FE (ie. not in
jobs or in rest api for instance) and if it can help the user sort out /
fix things (ie. we do add displayedErrorMessage for "Cannot create
multiple draft versions for the same workflow" or "Cannot delete
[field], please update the label identifier field first", but not
"Object metadata does not exist"), even if in practice in the FE users
should not be able to perform an action that will not work (ie should
not be able to save creation of multiple draft versions of the same
workflows).
**FE**
To ease the usage we replaced enqueueSnackBar with enqueueErrorSnackBar
and enqueueSuccessSnackBar with an api that only requires to pass on the
error.
If no displayedErrorMessage is specified then the default error message
is `An error occured.`
This PR aims at improving readability in sentry and user experience with
runtime errors.
**GraphQL errors (and ApolloError)**
1. In sentry we have a lot of "Object captured as exception with keys:
extensions, message" errors (2k over the last 90d), on which we have
zero information. This is because in apollo-factory we were passing on
GraphQL errors to sentry directly why sentry expects the structure of a
JS Error. We are now changing that, rebuilding an Error object and
attempting to help grouping by creating a fingerPrint based on error
code and truncated operationName (same as we do in the back for 500
graphql errors).
2. In sentry we have a lot of ApolloError, who actually correspond to
errors that should not be logged in sentry (Forbidden errors such as
"Email is not verified"), or errors that are already tracked by back-end
(Postgres errors such as "column xxx does not exist"). This is because
ApolloErrors become unhandled rejections errors if they are not caught
and automatically sent to sentry through the basic config. To change
that we are now filtering out ApolloErrors created from GraphQL Errors
before sending error to sentry:
<img width="524" alt="image"
src="https://github.com/user-attachments/assets/02974829-26d9-4a9e-8c4c-cfe70155e4ab"
/>
**Runtime errors**
4. Runtime errors were all caught by sentry with the name "Error",
making them not easy to differentiate on sentry (they were not grouped
together but all appeared in the list as "Error"). We are replacing the
"Error" name with the error message, or the error code if present. We
are introducing a CustomError class that allows errors whose message
contain dynamic text (an id for instance) to be identified on sentry
with a common code. _(TODO: if this approach is validated then I have
yet to replace Error with dynamic error messages with CustomError)_
5. Runtime error messages contain technical details that do not mean
anything to users (for instance, "Invalid folder ID: ${droppableId}",
"ObjectMetadataItem not found", etc.). Let's replace them with "Please
refresh the page." to users and keep the message error for sentry and
our dev experience (they will still show in the console as uncaught
errors).
# Introduction
This PR might have a lot of impact on tested validation
Avoid catching programmatically thrown error
---------
Co-authored-by: Charles Bochet <charles@twenty.com>
## 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"
/>
We were using a global ValidationPipe in main.ts. This is an issue as
@Controllers should return HttpExecption and @Resolvers should return
GraphqlErrors
Removing the global pipe and creating a ResolverValidationPipe able to
generate GraphqlError. We also need to handle the exception in a filter
to avoid nest to think it's unhandled and make it flow to logs
Next step:
- it would be nice to have both @UsePipes(ResolverValidationPipe) +
@UseFilters(GraphqlValidationExceptionFilter) come together. This should
be possible if we create a @GraphQLResolver annotation
Yoga graphql error were not correctly interpreted by the exception
handler. Mostly validations on the scalars such as bad enum options,
wrong format for uuid and such.
This PR adds a new convertGraphQLErrorToBaseGraphQLError utility
function in graphql-errors.util.ts that converts those errors to our
custom BaseGraphQLError by using the extension.http.code from the error
when possible so they can be handled the same way we treat the graphql
errors we throw ourselves.
Before
<img width="799" alt="Screenshot 2025-05-16 at 11 04 08"
src="https://github.com/user-attachments/assets/08b0a908-34d8-45a6-b315-8e211d1104ce"
/>
After
<img width="797" alt="Screenshot 2025-05-16 at 11 16 37"
src="https://github.com/user-attachments/assets/3fff0a70-6c3f-413a-b458-56030377fec9"
/>
Follow-up on https://github.com/twentyhq/twenty/pull/12007
In this PR
- adding a filter on HttpExceptionHandlerService to filter out 4xx
errors from driver handling (as we do for graphQL errors: see
useGraphQLErrorHandler hook - only filteredIssues are sent to`
exceptionHandlerService.captureExceptions()`.)
- grouping together more missing metadata issues
- attempting to use error codes as issues names in sentry to improve UI;
for now it says "Error" all the time
# Introduction
Added a no-explicit-any rule to the twenty-server, not applicable to
tests and integration tests folder
Related to https://github.com/twentyhq/core-team-issues/issues/975
Discussed with Charles
## In case of conflicts
Until this is approved I won't rebased and handle conflict, just need to
drop two latest commits and re run the scripts etc
## Legacy
We decided not to handle the existing lint error occurrences and
programmatically ignored them through a disable next line rule comment
## Open question
We might wanna activate the
[no-explicit-any](https://typescript-eslint.io/rules/no-explicit-any/)
`ignoreRestArgs` for our use case ?
```
ignoreRestArgs?: boolean;
```
---------
Co-authored-by: etiennejouan <jouan.etienne@gmail.com>
This PR attemps at improving sentry grouping and filtering by
- Using the exceptionCode as the fingerprint when the error is a
customException. For this to work in this PR we are now throwing
customExceptions instead of internalServerError deprived of their code.
They will still be converted to Internal server errors when sent back as
response
- Filtering 4xx issues where it was missing (for emailVerification
because errors were not handled, for invalid captcha and billing errors
because they are httpErrors and not graphqlErrors)
---------
Co-authored-by: Félix Malfait <felix@twenty.com>
Etienne