Custom metrics (#8)

* initial commit

* updates

* updates

* fix

* updates per PR review

Readme.md

@@ -20,6 +20,7 @@ A lot of these samples require an [Azure subscription, click here to get started
 | [Linux metrics with Telegraf and Azure Monitor](linux_metrics_telegraf_azuremonitor) | This script installs and configures [Telegraf](https://github.com/influxdata/telegraf) to send VM performance metrics (CPU/memory/disk/network) to [Azure Monitor](https://learn.microsoft.com/en-us/azure/azure-monitor/overview). |
 | [Linux logs with Fluent Bit and Azure Data Explorer (Kusto)](linux_logs_fluentbit_kusto) | This script installs and configures [Fluent Bit](https://github.com/fluent/fluent-bit) to capture real-time logs from your game servers and send them to [Azure Data Explorer (Kusto)](https://learn.microsoft.com/en-us/azure/data-explorer/data-explorer-overview). |
 | [Linux logs with Fluent Bit and Azure Blob Storage](linux_logs_fluentbit_azurestorage) | This script installs and configures [Fluent Bit](https://github.com/fluent/fluent-bit) to capture real-time logs from your game servers and send them to [Azure Blob Storage](https://learn.microsoft.com/en-us/azure/storage/blobs/storage-blobs-introduction). |
+| [Linux custom game server metrics](windows_linux_custom_gameserver_metrics) | This script shows you how to send metrics from your game server to [Telegraf](https://github.com/influxdata/telegraf) and then to the backend of your choice |
 
 ### Windows
 
@@ -30,6 +31,7 @@ A lot of these samples require an [Azure subscription, click here to get started
 | [Windows metrics with Telegraf and Application Insights](windows_metrics_telegraf_applicationinsights) | This script installs and configures [Telegraf](https://github.com/influxdata/telegraf) to send VM performance metrics (CPU/memory/disk/network) to [Application Insights](https://learn.microsoft.com/en-us/azure/azure-monitor/app/app-insights-overview). |
 | [Windows logs with Telegraf and PlayFab](windows_logs_telegraf_playfab) | This script installs and configures [Telegraf](https://github.com/influxdata/telegraf) to grab game server logs and send them to [PlayFab](https://learn.microsoft.com/en-us/gaming/playfab/). |
 | [Windows enable optional diagnostics](windows_enable_optional_diagnostics) | This script enables sending diagnostic data notify Microsoft of application faults, kernel faults, unresponsive applications, and other application specific problems. |
+| [Windows custom game server metrics](windows_linux_custom_gameserver_metrics) | This script shows you how to send metrics from your game server to [Telegraf](https://github.com/influxdata/telegraf) and then to the backend of your choice |
 
 ## Debugging
 

windows_linux_custom_gameserver_metrics/PF_StartupScript.ps1

@@ -0,0 +1,41 @@
+<#
+.SYNOPSIS
+  <Overview of script>
+.DESCRIPTION
+  <Brief description of script>
+.NOTES
+  Version:        1.0
+  Author:         <Name>
+  Creation Date:  <Date>
+  Purpose/Change: Initial script development
+#>
+
+# Set Error Action to Stop
+$ErrorActionPreference = "Stop"
+
+# Grab the script path 
+$scriptPath = split-path -parent $MyInvocation.MyCommand.Definition
+echo "the current script path is $scriptPath"
+
+# these are some of the env variables that are available to you
+echo "PlayFab Title ID is $env:PF_TITLE_ID" # e.g. 59F84
+echo "PlayFab Build ID is $env:PF_BUILD_ID" # Guid, e.g. 09d91059-22d3-4d0a-8c99-f34f80525aae
+echo "PlayFab Virtual Machine ID is $env:PF_VM_ID" # e.g. vmss:SouthCentralUs:2458795A9259968E_12fe54be-fae1-41aa-83d9-09b809d5ef01:09d91059-22d3-4d0a-8c99-f34f80525aae
+echo "Region where the VM is deployed is $env:PF_REGION" # e.g. SouthCentralUs
+echo "Shared content folder is $env:PF_SHARED_CONTENT_FOLDER_VM" # e.g. D:\sharedcontentfolder (All servers running on this VM have access to this folder through the PF_SHARED_CONTENT_FOLDER env variable.)
+
+# configure the telegraf configuration file with proper dimensions (title ID/build ID/vm ID)
+$telegrafConfPath = "$scriptPath\telegraf.conf"
+((Get-Content -path $telegrafConfPath -Raw) -replace '_%PF_TITLE_ID%_', "$env:PF_TITLE_ID") | Set-Content -Path $telegrafConfPath
+((Get-Content -path $telegrafConfPath -Raw) -replace '_%PF_BUILD_ID%_', "$env:PF_BUILD_ID") | Set-Content -Path $telegrafConfPath
+((Get-Content -path $telegrafConfPath -Raw) -replace '_%PF_VM_ID%_', "$env:PF_VM_ID") | Set-Content -Path $telegrafConfPath
+
+# install telegraf as a Windows service
+$telegrafPath = 'C:\Program Files\telegraf'
+New-Item -ItemType Directory -Force -Path "$telegrafPath"
+cp "$scriptPath\telegraf.*" "$telegrafPath"
+cd "$telegrafPath"
+.\telegraf.exe --service install --config "$telegrafPath\telegraf.conf"
+
+# start the telegraf service
+.\telegraf.exe --service start

windows_linux_custom_gameserver_metrics/PF_StartupScript.sh

@@ -0,0 +1,26 @@
+#!/bin/bash
+set -o errexit # script exits when a command fails == set -e
+set -o nounset # script exits when tries to use undeclared variables == set -u
+set -o xtrace # trace what's executed == set -x (useful for debugging)
+set -o pipefail # causes pipelines to retain / set the last non-zero status
+
+# get script's path
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
+
+# install telegraf
+sudo apt-get -yq --no-install-recommends install "${DIR}"/telegraf.deb
+
+systemctl stop telegraf
+
+# remove \r\n in-place
+sed -i -e 's/\r$//' ${DIR}/telegraf.conf 
+
+mkdir -p /etc/telegraf || echo 'Warning: /etc/telegraf already exists'
+
+# add PF_TITLE_ID, PF_BUILD_ID, PF_VM_ID to telegraf.conf as dimensions
+sed -e "s/_%PF_TITLE_ID%_/${PF_TITLE_ID}/g" -e "s/_%PF_BUILD_ID%_/${PF_BUILD_ID}/g" -e "s/_%PF_VM_ID%_/${PF_VM_ID}/g" -e "s/_%PF_REGION%_/${PF_REGION}/g" ${DIR}/telegraf.conf > /etc/telegraf/telegraf.conf
+
+# add user telegraf to group docker so telegraf daemon can read docker logs
+sudo usermod -a -G docker telegraf
+
+systemctl restart telegraf

windows_linux_custom_gameserver_metrics/README.md

@@ -0,0 +1,62 @@
+# Windows/Linux custom game server metrics
+
+## Introduction
+
+[Telegraf](https://github.com/influxdata/telegraf) is a plugin-driven server agent for collecting and reporting metrics. It supports a wide variety of inputs, processors, aggregators, and outputs.
+
+## What it does
+
+This script uses [telegraf](https://www.influxdata.com/time-series-platform/telegraf/) agent that sets up an HTTP server to listen for external metrics. Using this solution, you can send any kind of metrics from your game server to telegraf and you can use any of the [telegraf output plugins](https://github.com/influxdata/telegraf/blob/master/docs/OUTPUTS.md) to store them, in the backend of your choice.
+
+The documentation for the [http listener telegraf input plugin is here](https://github.com/influxdata/telegraf/blob/master/plugins/inputs/http_listener_v2/README.md), check it out for instructions on how to properly configure it. You can also check this [telegraf.conf sample config file](./telegraf.conf) file for an example of how to configure telegraf to listen for metrics on port 8080.
+
+To send custom metrics, you need send a POST request to the telegraf server from your game server. To do that, you need to find the host VM IP address. This is different depending on the type of Build you are using.
+
+### Host IP using Windows or Linux Processes
+
+If you are using [process mode](https://learn.microsoft.com/en-us/gaming/playfab/features/multiplayer/servers/deploy-using-game-manager#server-details-for-process-mode), you can talk to telegraf using `localhost` or `127.0.0.1`.
+
+### Host IP using Windows Containers
+
+PlayFab configures a bridge network between containers and the host VM. To get the IP of the host when using Windows Containers, you can use the following Powershell script from inside the container, before you start your game server:
+
+```powershell
+$gateway = (Get-WmiObject -Class Win32_IP4RouteTable | where { $_.destination -eq '0.0.0.0' -and $_.mask -eq '0.0.0.0' } | sort metric1 | select nexthop).nexthop
+Write-Output $gateway
+```
+
+### Host IP using Linux Containers
+
+PlayFab configures a bridge network between containers and the host VM. To get the IP of the host when using Linux Containers, you can run a script from inside the container, before you start your game server. The script you can run depends on the OS that you configure your container image with. For example, if you are using Ubuntu, you can do `sudo apt update && sudo apt install net-tools` to install `netstat` and then grab the host VM (which acts as a gateway) IP using:
+
+```bash
+netstat -rn | grep '^0.0.0.0' | awk '{print $2}'
+```
+
+## Sending metrics to telegraf
+
+To send metrics to telegraf, you'd need to send the data in a specific format. You can see the documentation [here](https://github.com/influxdata/telegraf/blob/master/plugins/inputs/http_listener_v2/README.md) and a couple of samples below:
+
+```bash
+curl -i -XPOST 'http://localhost:8080/telegraf' --data-binary 'cpu_load_short,host=server01,region=us-west value=0.64 1434055562000000000'
+```
+
+```powershell
+$headers = @{
+    "Content-Type" = "application/octet-stream"
+}
+$body = "cpu_load_short,host=server01,region=us-west value=0.64 1434055562000000000"
+$response = Invoke-WebRequest -Uri "http://localhost:8080/telegraf" -Method POST -Headers $headers -Body $body
+```
+
+## Usage
+
+You should download telegraf from the [GitHub releases](https://github.com/influxdata/telegraf/releases) section. The [Windows amd64 package](https://dl.influxdata.com/telegraf/releases/telegraf-1.24.4_windows_amd64.zip) works with this sample.
+
+Then, create a zip file with the following contents:
+
+- telegraf.conf
+- telegraf package file (e.g. telegraf.exe)
+- PF_StartupScript.ps1 or PF_StartupScript.sh file (depending on the OS you are using)
+
+You can now create a new MPS Build with your startup script using the [instructions here](https://learn.microsoft.com/en-us/gaming/playfab/features/multiplayer/servers/vmstartupscript).

windows_linux_custom_gameserver_metrics/telegraf.conf

@@ -0,0 +1,28 @@
+[agent]
+  interval = "10s" 
+  round_interval = true
+  metric_batch_size = 1000
+  metric_buffer_limit = 10000
+  collection_jitter = "0s"
+  flush_interval = "60s"
+  flush_jitter = "0s"
+  precision = ""
+  debug = true
+  omit_hostname = true
+
+[global_tags]
+  titleID = "_%PF_TITLE_ID%_"
+  buildID = "_%PF_BUILD_ID%_"
+  vmID = "_%PF_VM_ID%_"
+
+[[outputs.file]]
+  ## Files to write to, "stdout" is a specially handled file.
+  # this is just for testing, you should add the outputs of your choice
+  files = ["stdout"]
+
+# Generic HTTP write listener
+[[inputs.http_listener_v2]]
+  ## Address and port to host HTTP listener on
+  service_address = ":8080"
+  data_format = "influx"
+