Google Cloud Run Speedrun

Posted on In

1. setup Google Cloud Run

click Connect Repo

choose your repository and select Dockerfile; choose your git branch

use reasonable service name and region

select request-based billing

set auto-scaling from 0 to a small number (no need for 100), save budget

1.0 You need a working Dockerfile

template:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
# Use the official lightweight Python image
FROM python:3.11-slim

# Set the working directory to /app
WORKDIR /app

# Copy the requirements file into the container
COPY requirements.txt .

# Install the dependencies
RUN pip install --no-cache-dir -r requirements.txt

# Copy the rest of the application code
COPY . .

# Set environment variables
ENV PORT=8080
ENV PYTHONUNBUFFERED=1
ENV FLASK_APP=app.py

# Expose port 8080 for the application
EXPOSE 8080

# Use Gunicorn as the application server
# RUN python init.py
CMD ["gunicorn", "--bind", "0.0.0.0:8080", "app:app", "--log-level", "debug"]

2. trigger by tag instead of commit

2.0 the setup

click edit repo settings

in Event, select Push new tag

set Tag to ^v.*$, so that when a new tag like v1.0.1, v0.2.1 is tagged in the newest commit, build will be triggered

2.1 push with tags

1
2
git tag v1.0.0 # set tag to current commit or right click on current commit on VScode
git push --tags # push all tags

then a new build is triggered.

3. if you use google cloud storage

3.0 create a service account

goto IAM&Admin, left panel Service Account, click Create Service Account

give reasonable name

grant role of Storage Object Admin or other role you prefer

click on action ..., manage keys, and download the json key (keep it secure)