clarenceb

I read through docs starting here: https://bradygaster.github.io/squad/docs/get-started/installation/

I have some questions as squad is a bit confusing:

1. Seems squad is designed to run locally? There is copilot coding agent support but the docs seem to imply it adds the coding agent as another member to run tasks in the cloud as opposed to running all team agents concurrently in the cloud?
2. What's difference between running squad vs copilot --agent squad? The later seems to work better for me.
3. To customise your squad, do you need to edit all the underlying files that were created or can you prompt each squad member to adjust their prompt, skills, model, etc.?
4. After a squad completes, what if I want to work on a different task and perhaps form a different squad. Are all those squad files and members polluting my repo? Can I resume from a previous squad if I need to?

Let me research the Squad repository to answer your questions thoroughly.Great questions! After digging through the Squad rep

clarenceb

{
  "keys": [
    {
      "kty": "RSA",
      "use": "sig",
      "alg": "RS384",
      "kid": "61661a07bb6c8cb2bfd5d605a855b411",
      "n": "ryc4YAWwE8VlVXAI_2pcGPJVmZoI17yWl6MTi167sEclOPpLfZiuGKw0KYCEfrzFwVCm8J8UyQUD1_k514e7bOc1Q72lBY7U_dY3hMjp-Da1MseJsh_XCkHP8BXsUW4JKGt5_dOkPVyZ8pRX4I_mAXQ8elOmp78tjB3O4P80B-Vmwm3ZjceTnsGz2m8TQi-3OoFt-fnkfp-X4aGAx64DIQ2Wp4RTb7aBQ7G7CA67lisaYkkNqti49CQiV5or-skhu9LG85CoMGdis3tVxLyvvSfowhlD0dNweu9g9-oIRMtMl6yTuXLknxAHLrDsfKvL1npw4-31uCIKGB4xFZwSMw",
      "e": "AQAB"
    }

clarenceb

@description('Name of the Logic App')
param name string

@description('Add unique suffix to Logic App name for global uniqueness')
param addUniqueSuffix bool = true

@description('Subscription ID')
param subscriptionId string = subscription().subscriptionId

@description('Location for all resources')

clarenceb

az group create -n test1 -l australiaeast

# Create an AKS Automatic cluster
az aks create -name testcluster1 -g test1 --sku automatic

az aks show -n testcluster1 -g test1 | jq '.sku

#{
#  "name": "Automatic",
#  "tier": "Standard"

clarenceb

from locust import HttpUser, task, between

class UserBehavior(HttpUser):
    """
    Represents the user behavior for load testing.

    This class defines the tasks that the user will perform during the load testing.

    It is functionally similar to the JMeter test plan (SampleApp.jmx), where the user performs a series of tasks to simulate real-world
    """

clarenceb

import os
from locust import HttpUser, LoadTestShape, TaskSet, between, task

class UserTasks(TaskSet):
    """
    Represents the user tasks for load testing (not sequential).

    This class defines the tasks that the user will perform during the load testing.

    It is functionally similar to the JMeter test plan (SampleApp.jmx), where the user performs a series of tasks to simulate real-world

clarenceb

#!/bin/bash

TENANT="<your-tenant>.onmicrosoft.com"
SUBSCRIPTION_ID="<your-azure-subscription-id>"
BROWSER="/usr/bin/wslview"

export BROWSER

# Fix WSL2 clock synchronisation drift which commonly occurs
sudo ntpdate pool.ntp.org

clarenceb

I have found a way to build Windows container images without Docker - It doesn't use a Dockerfile though.

The way I got this to work is using crane from Google (https://github.com/google/go-containerregistry/blob/main/cmd/crane/doc/crane.md) which lets you append image layers to a base image. You can build Windows images on either Linux or Windows or inside of a container (Linux/Windows) - however, you'll need to build any artefacts (e.g. EXE, DLLs) on a Windows machine since this approach is using Crane to assemble a Windows container image on a Linux host (container).

Here is a quick PoC I did to prove it works (using Docker Desktop with Windows Container mode on my laptop):

dotnet --version
# 7.0.102

clarenceb

Note: currently, release annotations only work with the classic metrics explorer.

# Guidance/Steps:
- Enable Application Insights for the prod web app
- Ensure the App Insights Instrumentation Key is correct in the web app Application Settings, set APPINSIGHTS_INSTRUMENTATIONKEY (to be extra safe, also add Keys:ApplicationInsights:InstrumentationKey with the same value since the code looks that setting up)
- Restart the web app (prod) and navigate to a few URLs in the web app, check that you can see metrics (server response time) appears in App Insights (prod) – reduce time scale to 30 mins

## Add release annotations
- See: https://docs.microsoft.com/en-us/azure/azure-monitor/app/annotations
- Edit the prod stage in the release pipeline

clarenceb

Note: The Azure DevOps Pipelines release gates for Azure Monitor alerts appear to (currently) only support Classic Alerts.

# Guidance/Steps:
Pre-requisites (Lab 5) – app insights enable, server delay introduced

## Setup the alert
- From App Insights (prod), go to Configure / Alerts, then click “View classic alerts” from the top nav bar
- Click “Add metric alert (classic)”
- Enter name “slow-server”
- Select the prod App Insights instance