Skip to content

Commit b2d347b

Browse files
gheorghitahurmuzgheorghitahurmuz
authored
Merge pull request #371 from UiPath/chore/add-tracing-doc-nav
docs: add Tracing to nav
2 parents a48e902 + 50019c0 commit b2d347b

File tree

2 files changed

+121
-120
lines changed

2 files changed

+121
-120
lines changed

‎docs/core/traced.md

Lines changed: 120 additions & 120 deletions
Original file line numberDiff line numberDiff line change
@@ -1,121 +1,121 @@
1-
# Tracing
2-
3-
The `traced()` decorator enables automatic tracing of function calls, inputs, and outputs. It is designed to help you monitor, debug, and audit your code by capturing detailed information about function executions, including arguments, return values, and exceptions.
4-
5-
You can view the traces of an Orchestrator job by going to the Jobs page, click a job, a side panel will open, and they will be available under the `Trace` tab. These can also be seen in UiPath Maestro when your agent is part of a larger process orchestration.
6-
7-
## Usage
8-
9-
Apply the `@traced()` decorator to any function (sync, async, generator, or async generator) to automatically record its execution as a trace span.
10-
11-
```python
12-
from uipath.tracing._traced import traced
13-
14-
@traced()
15-
def my_function(x, y):
16-
return x + y
17-
18-
@traced(name="custom_span", run_type="my_type")
19-
async def my_async_function(a, b):
20-
return a * b
21-
```
22-
23-
## Parameters
24-
25-
| Parameter | Type | Description |
26-
|------------------|-------------------------------------|---------------------------------------------------------------------------------------------------|
27-
| name | Optional[str] | Custom name for the trace span. Defaults to the function name. |
28-
| run_type | Optional[str] | Category for the run (e.g., "uipath"). Useful for filtering traces. |
29-
| span_type | Optional[str] | Custom type for the span. Defaults to function type (sync/async/generator). |
30-
| input_processor | Optional[Callable[[dict], dict]] | Function to process/transform inputs before recording. Receives a dict of arguments. |
31-
| output_processor | Optional[Callable[[Any], Any]] | Function to process/transform outputs before recording. Receives the function's return value. |
32-
| hide_input | bool | If True, input data is redacted in the trace for privacy/security. |
33-
| hide_output | bool | If True, output data is redacted in the trace for privacy/security. |
34-
35-
## Input and Output Processors
36-
37-
Processors allow you to mask, redact, or transform sensitive data before it is recorded in the trace. For example:
38-
39-
```python
40-
def mask_inputs(inputs):
41-
inputs = inputs.copy()
42-
if 'password' in inputs:
43-
inputs['password'] = '***REDACTED***'
44-
return inputs
45-
46-
def anonymize_output(output):
47-
if isinstance(output, dict) and 'email' in output:
48-
output = output.copy()
49-
output['email'] = 'anonymous@example.com'
50-
return output
51-
52-
@traced(input_processor=mask_inputs, output_processor=anonymize_output)
53-
def login(user, password):
54-
# ...
55-
return {"email": user + "@example.com"}
56-
```
57-
58-
## Privacy Controls
59-
60-
- Set `hide_input=True` to prevent input data from being logged.
61-
- Set `hide_output=True` to prevent output data from being logged.
62-
63-
```python
64-
@traced(hide_input=True, hide_output=True)
65-
def sensitive_operation(secret):
66-
...
67-
```
68-
69-
## Supported Function Types
70-
71-
- Regular functions (sync/async)
72-
- Generator functions (sync/async)
73-
74-
## Example with plain python agents
75-
76-
When used with plain python agents please call `wait_for_tracers()` at the end of the script to ensure all traces are sent, if this is not called the agent could end without sending all the traces.
77-
78-
```python
79-
80-
from uipath.tracing import traced, wait_for_tracers
81-
82-
@traced(name="process_payment", run_type="payment", hide_input=True)
83-
def process_payment(card_number, amount):
84-
# Sensitive input will not be logged
85-
return {"status": "success", "amount": amount}
86-
87-
@traced()
88-
def main():
89-
process_payment()
90-
91-
def main_wait_traces():
92-
try:
93-
main()
94-
finally:
95-
# this needs to be called after the last `traced` function is done
96-
# to ensure the trace associated with main is saved
97-
wait_for_tracers()
98-
99-
if __name__ == "__main__":
100-
main_wait_traces()
101-
```
102-
103-
104-
## Example with langchain agents
105-
106-
When using `uipath-langchain` there is no need to call wait_for_tracers our framework will ensure that is called.
107-
108-
```python
109-
@traced()
110-
def my_custom_traced_function(input: str) -> str:
111-
return { "x": "some-output" }
112-
```
113-
114-
You can also use `@traceable()` attribute from langchain, but we recommend using `@traced()` attribute instead.
115-
116-
```python
117-
@traceable()
118-
# @traced() ---> do not use both at the same time or it will duplicate spans.
119-
def my_custom_traced_function(input: str) -> str:
120-
return { "x": "some-output" }
1+
# Tracing
2+
3+
The `traced()` decorator enables automatic tracing of function calls, inputs, and outputs. It is designed to help you monitor, debug, and audit your code by capturing detailed information about function executions, including arguments, return values, and exceptions.
4+
5+
You can view the traces of an Orchestrator job by going to the Jobs page, click a job, a side panel will open, and they will be available under the `Trace` tab. These can also be seen in UiPath Maestro when your agent is part of a larger process orchestration.
6+
7+
## Usage
8+
9+
Apply the `@traced()` decorator to any function (sync, async, generator, or async generator) to automatically record its execution as a trace span.
10+
11+
```python hl_lines="3 7"
12+
from uipath.tracing._traced import traced
13+
14+
@traced()
15+
def my_function(x, y):
16+
return x + y
17+
18+
@traced(name="custom_span", run_type="my_type")
19+
async def my_async_function(a, b):
20+
return a * b
21+
```
22+
23+
## Parameters
24+
25+
| Parameter | Type | Description |
26+
|------------------|-------------------------------------|---------------------------------------------------------------------------------------------------|
27+
| name | Optional[str] | Custom name for the trace span. Defaults to the function name. |
28+
| run_type | Optional[str] | Category for the run (e.g., "uipath"). Useful for filtering traces. |
29+
| span_type | Optional[str] | Custom type for the span. Defaults to function type (sync/async/generator). |
30+
| input_processor | Optional[Callable[[dict], dict]] | Function to process/transform inputs before recording. Receives a dict of arguments. |
31+
| output_processor | Optional[Callable[[Any], Any]] | Function to process/transform outputs before recording. Receives the function's return value. |
32+
| hide_input | bool | If True, input data is redacted in the trace for privacy/security. |
33+
| hide_output | bool | If True, output data is redacted in the trace for privacy/security. |
34+
35+
## Input and Output Processors
36+
37+
Processors allow you to mask, redact, or transform sensitive data before it is recorded in the trace. For example:
38+
39+
```python hl_lines="13"
40+
def mask_inputs(inputs):
41+
inputs = inputs.copy()
42+
if 'password' in inputs:
43+
inputs['password'] = '***REDACTED***'
44+
return inputs
45+
46+
def anonymize_output(output):
47+
if isinstance(output, dict) and 'email' in output:
48+
output = output.copy()
49+
output['email'] = 'anonymous@example.com'
50+
return output
51+
52+
@traced(input_processor=mask_inputs, output_processor=anonymize_output)
53+
def login(user, password):
54+
# ...
55+
return {"email": user + "@example.com"}
56+
```
57+
58+
## Privacy Controls
59+
60+
- Set `hide_input=True` to prevent input data from being logged.
61+
- Set `hide_output=True` to prevent output data from being logged.
62+
63+
```python hl_lines="1"
64+
@traced(hide_input=True, hide_output=True)
65+
def sensitive_operation(secret):
66+
...
67+
```
68+
69+
## Supported Function Types
70+
71+
- Regular functions (sync/async)
72+
- Generator functions (sync/async)
73+
74+
## Example with plain python agents
75+
76+
When used with plain python agents please call `wait_for_tracers()` at the end of the script to ensure all traces are sent, if this is not called the agent could end without sending all the traces.
77+
78+
```python hl_lines="3 8"
79+
80+
from uipath.tracing import traced, wait_for_tracers
81+
82+
@traced(name="process_payment", run_type="payment", hide_input=True)
83+
def process_payment(card_number, amount):
84+
# Sensitive input will not be logged
85+
return {"status": "success", "amount": amount}
86+
87+
@traced()
88+
def main():
89+
process_payment()
90+
91+
def main_wait_traces():
92+
try:
93+
main()
94+
finally:
95+
# this needs to be called after the last `traced` function is done
96+
# to ensure the trace associated with main is saved
97+
wait_for_tracers()
98+
99+
if __name__ == "__main__":
100+
main_wait_traces()
101+
```
102+
103+
104+
## Example with langchain agents
105+
106+
When using `uipath-langchain` there is no need to call wait_for_tracers our framework will ensure that is called.
107+
108+
```python hl_lines="1"
109+
@traced()
110+
def my_custom_traced_function(input: str) -> str:
111+
return { "x": "some-output" }
112+
```
113+
114+
You can also use `@traceable()` attribute from langchain, but we recommend using `@traced()` attribute instead.
115+
116+
```python hl_lines="1"
117+
@traceable()
118+
# @traced() ---> do not use both at the same time or it will duplicate spans.
119+
def my_custom_traced_function(input: str) -> str:
120+
return { "x": "some-output" }
121121
```

‎mkdocs.yml

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -61,6 +61,7 @@ nav:
6161
- Getting Started: core/getting_started.md
6262
- Environment Variables: core/environment_variables.md
6363
- CLI Reference: cli/index.md
64+
- Tracing: core/traced.md
6465
- Services:
6566
- Actions: core/actions.md
6667
- Assets: core/assets.md

0 commit comments

Comments
 (0)