Search⌘ K
AI Features
Home
Courses
Build a REST API Using Python and Deploy it to Microsoft Azure

Implement Logging in API Using Application Insights

Explore how to implement logging in your Python REST API using Application Insights on Microsoft Azure. Understand how to capture and log text data, sentiment, and key phrases effectively, and learn to integrate this into your code for improved API monitoring and diagnostics.

We'll cover the following...

Implement logging in our API

We have already set up the Application Insights resource and installed the required package to connect to the service for logging the data. Now, let’s make some changes in the code to do the logging operation.

We will log three things:

  • The text
  • The sentiment of the overall text
  • The key phrases from the overall text

Let us see the code now.

from fastapi import FastAPI
from pydantic import BaseModel
import logging
from opencensus.ext.azure.log_exporter import AzureLogHandler

import utils

app = FastAPI()

logger = logging.getLogger(__name__)
logger.setLevel(10)
# logger.addHandler(AzureLogHandler(connection_string='InstrumentationKey=<YOUR_KEY>'))

headers = {
    "Ocp-Apim-Subscription-Key": "e16626ed2cfd4d87ae51fffbee899079",
    "Content-Type": "application/json",
    "Accept": "application/json"
}

class Model(BaseModel):
    text_to_analyze: list

@ app.post("/")
def analyze_text(text: Model):
    response = {"sentiment": [], "keyphrases": []}
    no_of_text = len(text.text_to_analyze)
    tasks = []
    for i in range(no_of_text):
        document = {"documents": [
            {"id": i+1, "language": "en", "text": text.text_to_analyze[i]}]}
        sentiment = utils.call_text_analytics_api(
            headers, document, endpoint='sentiment')
        keyphrases = utils.call_text_analytics_api(
            headers, document, endpoint='keyPhrases')

        log_data = {
            "custom_dimensions":
            {
                "text": text.text_to_analyze[i],
                "text_sentiment": sentiment["documents"][0]["sentiment"],
                "text_keyphrases": keyphrases["documents"][0]["keyPhrases"]
            }
        }
        logger.info('Text Processed Succesfully', extra=log_data)

        response["sentiment"].append(sentiment["documents"][0])
        response["keyphrases"].append(keyphrases["documents"][0])
    return response
Logging using application insights

Explanation ...