Appearance
Error Tracking Tools
One tool, get-error-groups, for reading the error groups captured from your applications — the recurring exceptions and crashes kendo's error tracking groups together by fingerprint, each with an occurrence count, the latest message, and (on request) the latest stack trace.
It folds two questions into a single call: "what's broken in this one project?" (pass project_id) and "what's broken across everything I can see, in the last week?" (omit project_id, pass since_days).
Feature-gated
Error tracking is behind the error-tracking feature flag. If it isn't enabled for your tenant, the tool returns a clear "not available" message instead of data.
Permission Model
- Reading error groups requires Error Groups: Read permission
- When you pass
project_id, you must also have access to that project (team membership oraccessAllProjects) - When you omit
project_id, results are automatically scoped to the projects you can access — you never see groups from a project you aren't a member of - Project owners and admins bypass these checks
get-error-groups
Get, list, or retrieve error groups across all accessible projects or for a single project. Read-only and safe to call repeatedly.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
project_id | integer | No | Filter to one project by ID. Omit to survey every accessible project. |
environment | string | No | Filter by environment name, e.g. production, staging, development. |
since_days | integer | No | Only groups last seen within this many days (e.g. 7 = last week, 30 = last month). Omit for all retained groups. |
limit | integer | No | Maximum groups to return. Default 20, max 500. |
include_stack_trace | boolean | No | Include each group's full latest stack trace. Default false. |
Behavior
- Lean by default. Returns the 20 most recent groups unless you raise
limit(up to500). When more groups match than were returned, thetruncatedflag istrue— raiselimitor narrow the filters to see the rest. - Stack traces are omitted by default to keep responses small. Set
include_stack_trace=true— ideally together with aproject_idand a smalllimit— when you actually need to read the traces. - Ordered by most recent failure first (
last_seen_atdescending, then occurrence count). - Each group includes
id,project_id,environment,release,exception_class,latest_message,occurrence_count,first_seen_at, andlast_seen_at(pluslatest_stack_tracewhen requested). - Marked as
#[IsReadOnly]and#[IsIdempotent].
Response
json
{
"count": 2,
"truncated": false,
"limit": 20,
"error_groups": [
{
"id": 41,
"project_id": 1,
"environment": "production",
"release": "v1.4.0",
"exception_class": "Illuminate\\Database\\QueryException",
"latest_message": "SQLSTATE[40001]: Serialization failure: 1213 Deadlock found...",
"occurrence_count": 312,
"first_seen_at": "2026-06-01T09:14:22+00:00",
"last_seen_at": "2026-06-07T07:48:10+00:00"
}
]
}Example Prompts
What errors have my projects hit in the last 7 days?
Show me the production error groups for the kendo project, with stack traces.
What's the most frequent crash across everything I can access this month?
See Also
- Reports — User-submitted bug reports and feedback (distinct from automatically-captured error groups)
- Resources — Read project and issue data
- Overview & Setup — Connect your AI assistant to kendo