	In Browser
	StumbleUpon
	del.icio.us
	Google
	Google Buzz
	reddit
	LinkedIn

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > Patch Scheduling Agent

Patch Scheduling Agent

Author: Venkata Sudhakar

Applying security patches quickly is essential, but patching the payment gateway at 2 PM on a Friday during peak shopping hours is reckless. ShopMax India needs a patch scheduling agent that balances patch urgency against system criticality and business traffic patterns to produce a safe, prioritised deployment schedule.

This tutorial builds a Gemini ADK agent that takes a list of pending patches with severity ratings, checks each system's maintenance window availability, and generates a week-ahead patch deployment schedule with team assignments.

The below example shows a patch scheduling agent for ShopMax India in a business context.

import asyncio
from google.adk.agents import LlmAgent
from google.adk.tools import FunctionTool
from google.adk.runners import Runner
from google.adk.sessions import InMemorySessionService
from google.genai import types

# Pending patches
PATCHES = [
    {"id": "PATCH-001", "system": "Payment Gateway", "severity": "critical",
     "description": "OpenSSL vulnerability fix", "downtime_required": False},
    {"id": "PATCH-002", "system": "WMS Server", "severity": "high",
     "description": "Linux kernel security update", "downtime_required": True},
    {"id": "PATCH-003", "system": "Employee Portal", "severity": "medium",
     "description": "Apache web server update", "downtime_required": True},
    {"id": "PATCH-004", "system": "Marketing Website", "severity": "low",
     "description": "CMS plugin update", "downtime_required": False},
]

# Maintenance windows by system
MAINTENANCE_WINDOWS = {
    "Payment Gateway": "Sunday 02:00-04:00",
    "WMS Server":      "Saturday 22:00-02:00",
    "Employee Portal": "Saturday 22:00-02:00",
    "Marketing Website": "Any weeknight 23:00-01:00",
}

SEVERITY_DAYS = {"critical": 1, "high": 7, "medium": 14, "low": 30}

def generate_patch_schedule() -> dict:
    schedule = []
    for p in PATCHES:
        window = MAINTENANCE_WINDOWS.get(p["system"], "TBD - define maintenance window")
        max_days = SEVERITY_DAYS[p["severity"]]
        if p["severity"] == "critical":
            target = "ASAP - within 24 hours"
            team = "Security + Platform Engineering (on-call)"
        elif p["severity"] == "high":
            target = f"Next maintenance window: {window}"
            team = "Platform Engineering"
        else:
            target = f"Within {max_days} days - next scheduled {window}"
            team = "IT Operations"
        schedule.append({
            "patch_id": p["id"],
            "system": p["system"],
            "severity": p["severity"],
            "description": p["description"],
            "downtime_required": p["downtime_required"],
            "target_window": target,
            "assigned_team": team,
        })
    schedule.sort(key=lambda x: ["critical","high","medium","low"].index(x["severity"]))
    return {
        "total_patches": len(schedule),
        "critical_count": sum(1 for p in schedule if p["severity"] == "critical"),
        "schedule": schedule,
    }

patch_tool = FunctionTool(func=generate_patch_schedule)

agent = LlmAgent(
    name="patch_scheduling_agent",
    model="gemini-2.0-flash",
    description="Generates a patch deployment schedule for ShopMax India systems by severity and maintenance window.",
    instruction=(
        "You are a patch management agent for ShopMax India. "
        "When asked for a patch schedule, call generate_patch_schedule. "
        "Present critical patches first with urgency. "
        "List the target window, assigned team, and whether downtime is required for each patch."
    ),
    tools=[patch_tool],
)

async def main():
    session_service = InMemorySessionService()
    runner = Runner(agent=agent, app_name="patch_app", session_service=session_service)
    session = await session_service.create_session(app_name="patch_app", user_id="it1")
    msg = types.Content(role="user", parts=[types.Part(
        text="Generate this week patch deployment schedule"
    )])
    async for event in runner.run_async(user_id="it1", session_id=session.id, new_message=msg):
        if event.is_final_response() and event.content:
            print(event.content.parts[0].text)

asyncio.run(main())

It gives the following output,

Patch Deployment Schedule - ShopMax India | Week of Apr 8, 2026

Total: 4 patches | 1 Critical

[CRITICAL] PATCH-001 - Payment Gateway
  Desc    : OpenSSL vulnerability fix
  Window  : ASAP - within 24 hours
  Team    : Security + Platform Engineering (on-call)
  Downtime: Not required (rolling patch)

[HIGH] PATCH-002 - WMS Server
  Desc    : Linux kernel security update
  Window  : Saturday 22:00-02:00
  Team    : Platform Engineering
  Downtime: Required - schedule warehouse ops pause

[MEDIUM] PATCH-003 - Employee Portal
  Desc    : Apache web server update
  Window  : Within 14 days - Saturday 22:00-02:00
  Team    : IT Operations
  Downtime: Required

[LOW] PATCH-004 - Marketing Website
  Desc    : CMS plugin update
  Window  : Within 30 days - Any weeknight 23:00-01:00
  Team    : IT Operations
  Downtime: Not required

ShopMax India can run this agent every Monday to generate the week's patch deployment schedule, automatically creating calendar invites for the assigned teams and Jira sub-tasks for each patch. Critical patches trigger an immediate Slack notification to the security team lead with the 24-hour deadline clearly stated.

Send your comments, suggestions or queries regarding this site to [email protected].