Idiomatic REST API: Telemetry (#786)

# What does this PR do?

Changes Telemetry API to follow more idiomatic REST


- [ ] Addresses issue (#issue)


## Test Plan

TBD, once i get an approval for rest endpoints
This commit is contained in:
Dinesh Yeduguru 2025-01-16 12:08:46 -08:00 committed by GitHub
parent c79b087552
commit 59eeaf7f81
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
6 changed files with 955 additions and 669 deletions

File diff suppressed because it is too large Load diff

View file

@ -836,10 +836,7 @@ components:
oneOf:
- $ref: '#/components/schemas/BenchmarkEvalTaskConfig'
- $ref: '#/components/schemas/AppEvalTaskConfig'
task_id:
type: string
required:
- task_id
- input_rows
- scoring_functions
- task_config
@ -1003,17 +1000,6 @@ components:
required:
- job_id
type: object
JobCancelRequest:
additionalProperties: false
properties:
job_id:
type: string
task_id:
type: string
required:
- job_id
- task_id
type: object
JobStatus:
enum:
- completed
@ -1178,6 +1164,16 @@ components:
required:
- data
type: object
ListRoutesResponse:
additionalProperties: false
properties:
data:
items:
$ref: '#/components/schemas/RouteInfo'
type: array
required:
- data
type: object
ListScoringFunctionsResponse:
additionalProperties: false
properties:
@ -1781,52 +1777,35 @@ components:
- chunks
- scores
type: object
QuerySpanTreeRequest:
QuerySpanTreeResponse:
additionalProperties: false
properties:
attributes_to_return:
items:
type: string
type: array
max_depth:
type: integer
span_id:
type: string
required:
- span_id
data:
additionalProperties:
$ref: '#/components/schemas/SpanWithStatus'
type: object
QuerySpansRequest:
additionalProperties: false
properties:
attribute_filters:
items:
$ref: '#/components/schemas/QueryCondition'
type: array
attributes_to_return:
items:
type: string
type: array
max_depth:
type: integer
required:
- attribute_filters
- attributes_to_return
- data
type: object
QueryTracesRequest:
QuerySpansResponse:
additionalProperties: false
properties:
attribute_filters:
data:
items:
$ref: '#/components/schemas/QueryCondition'
$ref: '#/components/schemas/Span'
type: array
limit:
type: integer
offset:
type: integer
order_by:
required:
- data
type: object
QueryTracesResponse:
additionalProperties: false
properties:
data:
items:
type: string
$ref: '#/components/schemas/Trace'
type: array
required:
- data
type: object
RegexParserScoringFnParams:
additionalProperties: false
@ -2082,10 +2061,7 @@ components:
oneOf:
- $ref: '#/components/schemas/BenchmarkEvalTaskConfig'
- $ref: '#/components/schemas/AppEvalTaskConfig'
task_id:
type: string
required:
- task_id
- task_config
type: object
RunShieldRequest:
@ -3916,9 +3892,14 @@ paths:
description: OK
tags:
- EvalTasks
/v1/eval/evaluate-rows:
/v1/eval/tasks/{task_id}/evaluations:
post:
parameters:
- in: path
name: task_id
required: true
schema:
type: string
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
@ -3948,9 +3929,14 @@ paths:
description: OK
tags:
- Eval
/v1/eval/jobs/cancel:
/v1/eval/tasks/{task_id}/jobs:
post:
parameters:
- in: path
name: task_id
required: true
schema:
type: string
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
@ -3969,26 +3955,61 @@ paths:
content:
application/json:
schema:
$ref: '#/components/schemas/JobCancelRequest'
$ref: '#/components/schemas/RunEvalRequest'
required: true
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/Job'
description: OK
tags:
- Eval
/v1/eval/jobs/{job_id}:
get:
/v1/eval/tasks/{task_id}/jobs/{job_id}:
delete:
parameters:
- in: path
name: task_id
required: true
schema:
type: string
- in: path
name: job_id
required: true
schema:
type: string
- in: query
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
description: OK
tags:
- Eval
get:
parameters:
- in: path
name: task_id
required: true
schema:
type: string
- in: path
name: job_id
required: true
schema:
type: string
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
@ -4014,7 +4035,7 @@ paths:
description: OK
tags:
- Eval
/v1/eval/jobs/{job_id}/result:
/v1/eval/tasks/{task_id}/jobs/{job_id}/result:
get:
parameters:
- in: path
@ -4022,7 +4043,7 @@ paths:
required: true
schema:
type: string
- in: query
- in: path
name: task_id
required: true
schema:
@ -4050,38 +4071,6 @@ paths:
description: OK
tags:
- Eval
/v1/eval/run:
post:
parameters:
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/RunEvalRequest'
required: true
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/Job'
description: OK
tags:
- Eval
/v1/health:
get:
parameters:
@ -4208,6 +4197,58 @@ paths:
description: OK
tags:
- Inference
/v1/inspect/providers:
get:
parameters:
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/ListProvidersResponse'
description: OK
tags:
- Inspect
/v1/inspect/routes:
get:
parameters:
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/ListRoutesResponse'
description: OK
tags:
- Inspect
/v1/memory-banks:
get:
parameters:
@ -4688,62 +4729,6 @@ paths:
description: OK
tags:
- PostTraining (Coming Soon)
/v1/providers/list:
get:
parameters:
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/ListProvidersResponse'
description: OK
tags:
- Inspect
/v1/routes/list:
get:
parameters:
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
content:
application/json:
schema:
additionalProperties:
items:
$ref: '#/components/schemas/RouteInfo'
type: array
type: object
description: OK
tags:
- Inspect
/v1/safety/run-shield:
post:
parameters:
@ -5048,7 +5033,7 @@ paths:
description: OK
tags:
- SyntheticDataGeneration (Coming Soon)
/v1/telemetry/log-event:
/v1/telemetry/events:
post:
parameters:
- description: JSON-encoded provider data which will be made available to the
@ -5076,9 +5061,28 @@ paths:
description: OK
tags:
- Telemetry
/v1/telemetry/query-span-tree:
post:
/v1/telemetry/spans:
get:
parameters:
- in: query
name: attribute_filters
required: true
schema:
items:
$ref: '#/components/schemas/QueryCondition'
type: array
- in: query
name: attributes_to_return
required: true
schema:
items:
type: string
type: array
- in: query
name: max_depth
required: false
schema:
type: integer
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
@ -5093,88 +5097,16 @@ paths:
required: false
schema:
type: string
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/QuerySpanTreeRequest'
required: true
responses:
'200':
content:
application/json:
schema:
additionalProperties:
$ref: '#/components/schemas/SpanWithStatus'
type: object
$ref: '#/components/schemas/QuerySpansResponse'
description: OK
tags:
- Telemetry
/v1/telemetry/query-spans:
post:
parameters:
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/QuerySpansRequest'
required: true
responses:
'200':
content:
application/jsonl:
schema:
$ref: '#/components/schemas/Span'
description: OK
tags:
- Telemetry
/v1/telemetry/query-traces:
post:
parameters:
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/QueryTracesRequest'
required: true
responses:
'200':
content:
application/jsonl:
schema:
$ref: '#/components/schemas/Trace'
description: OK
tags:
- Telemetry
/v1/telemetry/save-spans-to-dataset:
/v1/telemetry/spans/export:
post:
parameters:
- description: JSON-encoded provider data which will be made available to the
@ -5202,6 +5134,166 @@ paths:
description: OK
tags:
- Telemetry
/v1/telemetry/spans/{span_id}/tree:
get:
parameters:
- in: path
name: span_id
required: true
schema:
type: string
- in: query
name: attributes_to_return
required: false
schema:
items:
type: string
type: array
- in: query
name: max_depth
required: false
schema:
type: integer
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/QuerySpanTreeResponse'
description: OK
tags:
- Telemetry
/v1/telemetry/traces:
get:
parameters:
- in: query
name: attribute_filters
required: false
schema:
items:
$ref: '#/components/schemas/QueryCondition'
type: array
- in: query
name: limit
required: false
schema:
type: integer
- in: query
name: offset
required: false
schema:
type: integer
- in: query
name: order_by
required: false
schema:
items:
type: string
type: array
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/QueryTracesResponse'
description: OK
tags:
- Telemetry
/v1/telemetry/traces/{trace_id}:
get:
parameters:
- in: path
name: trace_id
required: true
schema:
type: string
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/Trace'
description: OK
tags:
- Telemetry
/v1/telemetry/traces/{trace_id}/spans/{span_id}:
get:
parameters:
- in: path
name: trace_id
required: true
schema:
type: string
- in: path
name: span_id
required: true
schema:
type: string
- description: JSON-encoded provider data which will be made available to the
adapter servicing the API
in: header
name: X-LlamaStack-Provider-Data
required: false
schema:
type: string
- description: Version of the client making the request. This is used to ensure
that the client and server are compatible.
in: header
name: X-LlamaStack-Client-Version
required: false
schema:
type: string
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/Span'
description: OK
tags:
- Telemetry
/v1/tool-runtime/invoke:
post:
parameters:
@ -5673,9 +5765,6 @@ tags:
name: InvokeToolRequest
- description: <SchemaDefinition schemaRef="#/components/schemas/Job" />
name: Job
- description: <SchemaDefinition schemaRef="#/components/schemas/JobCancelRequest"
/>
name: JobCancelRequest
- description: <SchemaDefinition schemaRef="#/components/schemas/JobStatus" />
name: JobStatus
- description: <SchemaDefinition schemaRef="#/components/schemas/KeyValueMemoryBank"
@ -5711,6 +5800,9 @@ tags:
- description: <SchemaDefinition schemaRef="#/components/schemas/ListProvidersResponse"
/>
name: ListProvidersResponse
- description: <SchemaDefinition schemaRef="#/components/schemas/ListRoutesResponse"
/>
name: ListRoutesResponse
- description: <SchemaDefinition schemaRef="#/components/schemas/ListScoringFunctionsResponse"
/>
name: ListScoringFunctionsResponse
@ -5797,15 +5889,15 @@ tags:
- description: <SchemaDefinition schemaRef="#/components/schemas/QueryDocumentsResponse"
/>
name: QueryDocumentsResponse
- description: <SchemaDefinition schemaRef="#/components/schemas/QuerySpanTreeRequest"
- description: <SchemaDefinition schemaRef="#/components/schemas/QuerySpanTreeResponse"
/>
name: QuerySpanTreeRequest
- description: <SchemaDefinition schemaRef="#/components/schemas/QuerySpansRequest"
name: QuerySpanTreeResponse
- description: <SchemaDefinition schemaRef="#/components/schemas/QuerySpansResponse"
/>
name: QuerySpansRequest
- description: <SchemaDefinition schemaRef="#/components/schemas/QueryTracesRequest"
name: QuerySpansResponse
- description: <SchemaDefinition schemaRef="#/components/schemas/QueryTracesResponse"
/>
name: QueryTracesRequest
name: QueryTracesResponse
- description: <SchemaDefinition schemaRef="#/components/schemas/RegexParserScoringFnParams"
/>
name: RegexParserScoringFnParams
@ -6085,7 +6177,6 @@ x-tagGroups:
- InterleavedContentItem
- InvokeToolRequest
- Job
- JobCancelRequest
- JobStatus
- KeyValueMemoryBank
- KeyValueMemoryBankParams
@ -6098,6 +6189,7 @@ x-tagGroups:
- ListModelsResponse
- ListPostTrainingJobsResponse
- ListProvidersResponse
- ListRoutesResponse
- ListScoringFunctionsResponse
- ListShieldsResponse
- ListToolGroupsResponse
@ -6127,9 +6219,9 @@ x-tagGroups:
- QueryConditionOp
- QueryDocumentsRequest
- QueryDocumentsResponse
- QuerySpanTreeRequest
- QuerySpansRequest
- QueryTracesRequest
- QuerySpanTreeResponse
- QuerySpansResponse
- QueryTracesResponse
- RegexParserScoringFnParams
- RegisterDatasetRequest
- RegisterEvalTaskRequest

View file

@ -169,39 +169,57 @@ class QueryCondition(BaseModel):
value: Any
class QueryTracesResponse(BaseModel):
data: List[Trace]
class QuerySpansResponse(BaseModel):
data: List[Span]
class QuerySpanTreeResponse(BaseModel):
data: Dict[str, SpanWithStatus]
@runtime_checkable
class Telemetry(Protocol):
@webmethod(route="/telemetry/log-event")
@webmethod(route="/telemetry/events", method="POST")
async def log_event(
self, event: Event, ttl_seconds: int = DEFAULT_TTL_DAYS * 86400
) -> None: ...
@webmethod(route="/telemetry/query-traces", method="POST")
@webmethod(route="/telemetry/traces", method="GET")
async def query_traces(
self,
attribute_filters: Optional[List[QueryCondition]] = None,
limit: Optional[int] = 100,
offset: Optional[int] = 0,
order_by: Optional[List[str]] = None,
) -> List[Trace]: ...
) -> QueryTracesResponse: ...
@webmethod(route="/telemetry/query-span-tree", method="POST")
async def query_span_tree(
@webmethod(route="/telemetry/traces/{trace_id}", method="GET")
async def get_trace(self, trace_id: str) -> Trace: ...
@webmethod(route="/telemetry/traces/{trace_id}/spans/{span_id}", method="GET")
async def get_span(self, trace_id: str, span_id: str) -> Span: ...
@webmethod(route="/telemetry/spans/{span_id}/tree", method="GET")
async def get_span_tree(
self,
span_id: str,
attributes_to_return: Optional[List[str]] = None,
max_depth: Optional[int] = None,
) -> Dict[str, SpanWithStatus]: ...
) -> QuerySpanTreeResponse: ...
@webmethod(route="/telemetry/query-spans", method="POST")
@webmethod(route="/telemetry/spans", method="GET")
async def query_spans(
self,
attribute_filters: List[QueryCondition],
attributes_to_return: List[str],
max_depth: Optional[int] = None,
) -> List[Span]: ...
) -> QuerySpansResponse: ...
@webmethod(route="/telemetry/save-spans-to-dataset", method="POST")
@webmethod(route="/telemetry/spans/export", method="POST")
async def save_spans_to_dataset(
self,
attribute_filters: List[QueryCondition],

View file

@ -21,10 +21,12 @@ from llama_stack.apis.telemetry import (
Event,
MetricEvent,
QueryCondition,
QuerySpanTreeResponse,
QueryTracesResponse,
Span,
SpanEndPayload,
SpanStartPayload,
SpanStatus,
SpanWithStatus,
StructuredLogEvent,
Telemetry,
Trace,
@ -241,22 +243,32 @@ class TelemetryAdapter(TelemetryDatasetMixin, Telemetry):
limit: Optional[int] = 100,
offset: Optional[int] = 0,
order_by: Optional[List[str]] = None,
) -> List[Trace]:
return await self.trace_store.query_traces(
) -> QueryTracesResponse:
return QueryTracesResponse(
data=await self.trace_store.query_traces(
attribute_filters=attribute_filters,
limit=limit,
offset=offset,
order_by=order_by,
)
)
async def query_span_tree(
async def get_trace(self, trace_id: str) -> Trace:
return await self.trace_store.get_trace(trace_id)
async def get_span(self, trace_id: str, span_id: str) -> Span:
return await self.trace_store.get_span(trace_id, span_id)
async def get_span_tree(
self,
span_id: str,
attributes_to_return: Optional[List[str]] = None,
max_depth: Optional[int] = None,
) -> Dict[str, SpanWithStatus]:
return await self.trace_store.get_span_tree(
) -> QuerySpanTreeResponse:
return QuerySpanTreeResponse(
data=await self.trace_store.get_span_tree(
span_id=span_id,
attributes_to_return=attributes_to_return,
max_depth=max_depth,
)
)

View file

@ -7,7 +7,7 @@
from typing import List, Optional
from llama_stack.apis.datasetio import DatasetIO
from llama_stack.apis.telemetry import QueryCondition, Span
from llama_stack.apis.telemetry import QueryCondition, QuerySpansResponse, Span
class TelemetryDatasetMixin:
@ -48,18 +48,18 @@ class TelemetryDatasetMixin:
attribute_filters: List[QueryCondition],
attributes_to_return: List[str],
max_depth: Optional[int] = None,
) -> List[Span]:
) -> QuerySpansResponse:
traces = await self.query_traces(attribute_filters=attribute_filters)
spans = []
for trace in traces:
spans_by_id = await self.query_span_tree(
for trace in traces.data:
spans_by_id_resp = await self.get_span_tree(
span_id=trace.root_span_id,
attributes_to_return=attributes_to_return,
max_depth=max_depth,
)
for span in spans_by_id.values():
for span in spans_by_id_resp.data.values():
if span.attributes and all(
attr in span.attributes and span.attributes[attr] is not None
for attr in attributes_to_return
@ -76,4 +76,4 @@ class TelemetryDatasetMixin:
)
)
return spans
return QuerySpansResponse(data=spans)

View file

@ -10,7 +10,7 @@ from typing import Dict, List, Optional, Protocol
import aiosqlite
from llama_stack.apis.telemetry import QueryCondition, SpanWithStatus, Trace
from llama_stack.apis.telemetry import QueryCondition, Span, SpanWithStatus, Trace
class TraceStore(Protocol):
@ -167,3 +167,23 @@ class SQLiteTraceStore(TraceStore):
spans_by_id[span.span_id] = span
return spans_by_id
async def get_trace(self, trace_id: str) -> Trace:
query = "SELECT * FROM traces WHERE trace_id = ?"
async with aiosqlite.connect(self.conn_string) as conn:
conn.row_factory = aiosqlite.Row
async with conn.execute(query, (trace_id,)) as cursor:
row = await cursor.fetchone()
if row is None:
raise ValueError(f"Trace {trace_id} not found")
return Trace(**row)
async def get_span(self, trace_id: str, span_id: str) -> Span:
query = "SELECT * FROM spans WHERE trace_id = ? AND span_id = ?"
async with aiosqlite.connect(self.conn_string) as conn:
conn.row_factory = aiosqlite.Row
async with conn.execute(query, (trace_id, span_id)) as cursor:
row = await cursor.fetchone()
if row is None:
raise ValueError(f"Span {span_id} not found")
return Span(**row)