Automate Slack Notifications with Graphs Using Cloud Run Functions and Cloud Scheduler

I recently built a system to automate Slack notifications with graphs that visualize the session counts for the last 7 days. This was achieved using a combination of Cloud Run functions for data processing and graph generation, and Cloud Scheduler for scheduling the execution.

Implementation Overview

Cloud Run Function

The Cloud Run function queries BigQuery to fetch session data, uses Matplotlib to create a line chart, and then sends the chart to Slack via the Slack API. The following steps outline the setup process.

Here is the code for main.py. Before running, you need to set the SLACK_API_TOKEN and SLACK_CHANNEL_ID as environment variables. You can leave them empty for now, as we will set them up later.

import os
import matplotlib.pyplot as plt
from google.cloud import bigquery
from datetime import datetime, timedelta
import io
import pytz
from slack_sdk import WebClient
from slack_sdk.errors import SlackApiError

def create_weekly_total_sessions_chart(_):
    SLACK_TOKEN = os.environ.get('SLACK_API_TOKEN')
    SLACK_CHANNEL_ID = os.environ.get('SLACK_CHANNEL_ID')

    client = bigquery.Client()

    # Calculate the date range for the last 7 days
    jst = pytz.timezone('Asia/Tokyo')
    today = datetime.now(jst)
    start_date = (today - timedelta(days=7)).strftime('%Y-%m-%d')
    end_date = (today - timedelta(days=1)).strftime('%Y-%m-%d')

    query = f"""
        SELECT 
            DATE(created_at) AS date,
            COUNT(DISTINCT session_id) AS unique_sessions
        FROM `<project>.<dataset>.summary_all`
        WHERE created_at BETWEEN '{start_date} 00:00:00' AND '{end_date} 23:59:59'
        GROUP BY date
        ORDER BY date;
    """

    query_job = client.query(query)
    results = query_job.result()

    # Prepare data for the graph
    dates = []
    session_counts = []
    for row in results:
        dates.append(row['date'].strftime('%Y-%m-%d'))
        session_counts.append(row['unique_sessions'])

    # Generate the graph
    plt.figure()
    plt.plot(dates, session_counts, marker='o')
    plt.title('Unique Session Counts (Last 7 Days)')
    plt.xlabel('Date')
    plt.ylabel('Unique Sessions')
    plt.xticks(rotation=45)
    plt.tight_layout()

    # Save the graph as an image
    image_binary = io.BytesIO()
    plt.savefig(image_binary, format='png')
    image_binary.seek(0)

    # Send the graph to Slack
    client = WebClient(token=SLACK_TOKEN)
    try:
        response = client.files_upload_v2(
            channel=SLACK_CHANNEL_ID,
            file_uploads=[{
                "file": image_binary,
                "filename": "unique_sessions.png",
                "title": "Unique Session Counts (Last 7 Days)"
            }],
            initial_comment="Here are the session counts for the last 7 days!"
        )
    except SlackApiError as e:
        return f"Error uploading file: {e.response['error']}"

    return "Success"

Dependencies

Create a requirements.txt file and include the following dependencies:

functions-framework==3.*
google-cloud-bigquery
matplotlib
slack_sdk
pytz

Granting Access to the Cloud Run Function

To allow Cloud Scheduler or other services to invoke your Cloud Run function, you need to assign the roles/run.invoker role to the appropriate entity. Use the following command to do this:

gcloud functions add-invoker-policy-binding create-weekly-total-sessions-chart \
      --region="asia-northeast1" \
      --member="MEMBER_NAME"

Replace MEMBER_NAME with one of the following:

A service account for Cloud Scheduler: serviceAccount:scheduler-account@example.iam.gserviceaccount.com
For public access (not recommended): allUsers

Setting Up Cloud Scheduler

Use Cloud Scheduler to automate the function’s execution every Monday at 10:00 AM (JST). Here’s an example configuration:

Slack API Configuration

To enable your Cloud Run function to send Slack notifications, follow these steps:

1.Go to Slack API and create a new app.

2.Assign the following Bot Token Scopes under OAuth & Permissions:

channels:read
chat:write
files:write

3.Install the app to your Slack workspace and copy the Bot User OAuth Token.

4.Add the app to the Slack channel where you want to post notifications.

5.Copy the channel ID and paste it, along with the Bot Token, into the SLACK_CHANNEL_ID and SLACK_API_TOKEN environment variables for your Cloud Run function.

Final Result

Once everything is set up, your Slack channel will receive a weekly notification with a graph like this:

Blog