AccelByte Multiplayer Servers (AMS) Dashboards

Overview

The AccelByte Multiplayer Servers (AMS) dashboards in Grafana Cloud offer a clear view of important metrics that affect dedicated server performance. These dashboards gather data from AMS Services, providing real-time insights into key areas such as dedicated server resource usage and availability activity.

備考

To learn how to access and view Grafana Cloud dashboards, refer to the Access Grafana Cloud and View dashboards pages.

AMS dashboards

The AMS dashboards provided by AMS and Grafana Cloud integration are as follows:

• AMS Buffer Sizing
• AMS Cost Usage
• AMS DS Metrics
• AMS Fleet Overview

AMS buffer sizing

The AMS buffer sizing dashboard provides a way to understand and estimate your fleets buffer configuration. Data on this dashboard is based on historical usage. The AMS buffer sizing dashboard is divided into different sections composed of Global and Region. You can filter based on Environment, Fleet Identifier, Region, and AGS Namespace.

Global

This is the global summary of your claimed servers.

Panel	Description
Servers (Claimed)	The Total number of servers claimed across your fleet broken down by region and fleet.

Buffer in region

This is a regional break down, with suggested buffer configuration in your fleet, based on historical usage.

Panel	Description
Required Buffer [region: x, fleet: XYZ]	This is a summary of your required Buffer configurations, based on historical claim rates. Yellow indicates your Require buffer over 12 hours, and green indicates your short term buffers required. We recommend using this graph to model and adjust your buffer configurations as required.
Claimed Servers [region]	This is the actual number of server claim requests in that region at that time.
10 Minute Buffer required (Long-term trend) [region]	This is the 10 Minute Peak required buffers based on historical claim rate looking back on the last 12 hour, 10 hour, 8 hour, 6 hour periods along with the MAX over the time period of the graph
10 Minute Buffer required (short-term jitter) [region]	This is the peak 10 minute required buffers based on claim rate for the graph time period

AMS cost usage

The AMS cost usage dashboard is provides a summary of your spending on AMS. This dashboard should only be considered an estimate of your spend. The Dashboard is composed of two sections, one based on your environment spend and another based on your AMS account.

Total cost current environment

Total cost based on the associated AGS environment and selected AMS account

Panel	Description
Total Cost per Environment	Total cost of your environment for up to one year, the total cost for the previous month, and the total cost for this week.
All Time Cost Per Namespace($)	This is the total cost broken down by namespace, going back up to one year of AMS usage
All Time VM Cost ($)	This is the total cost broken down by type, going back up to one year of AMS usage
All Time Egress Cost ($)	This is the total cost of egress traffic, going back up to one year of AMS usage
Monthly Cost ($)	This is the total monthly cost

Total cost AMS account

Total costs as measured across your AMS account.

Panel	Description
Total Cost per AMS Account	Total cost all Time, Total cost for the past month, and Total for the past week
Monthly Cost ($)	Totals broken down on a month by month basis
VM Cost Weekly ($)	Totals broken down on a weekly basis
Egress Cost Weekly ($)	Total egress cost on a weekly basis

AMS DS metrics

The AMS DS Metrics dashboard is the dashboard to use to monitor the resource usage of dedicated servers on the underlying hosts. This dashboard provides valuable insight in to how much CPU, memory, and network traffic your dedicated servers are using. The dashboard breaks down usage by region.

Filters can be applied for Environment, Namespace, AMS Fleet, Region, and DS State.

Panel	Description
DS Max CPU Usage	This is the maximum CPU averaged across all DS processes in the fleet
DS Max Memory Usage	The average Max memory usage for all DS in that region and fleet
vCPU Idle per VM	The idle vCPU on each running VM
Free Memory per VM	The Free Memory on each running VM
Network Received per VM	Network receive traffic per VM
Network Transmited per VM	Network egress traffic per VM

AMS Fleet Overview

This AMS Fleet Overview dashboard is a summary of the state of your dedicated servers. This dashboard has three main sections, Overview, All Regions, and by region.

The Overview section provides a summary of metrics that are relevant when looking at an entire fleet.

Overview

Panel	Description
DS Creation to Ready duration	This provides the Max and P95 value for start time of the DS process. Allowing you to measure the start time of your DS process.
DS Claimed to Terminated duration	This provides the Max and P95 values for how long a claimed DS takes to transition to terminated state. This can be used to figure out your session lengths.
DS Crash Rate	The number of DS crashed across your fleet, by fleet ID and namespace

All Regions

The All regions section allows you to view a summary of all DS across all regions.

Panel	Description
Running DS	The number of concurrent running DS per fleet and state
Running DS (Stacked View)	The number of concurrent running DS per fleet and the status displayed in a stack view

Regional

The regions section allows you to view the state of your DS in your fleet on a region by region basis, along with the claim rates in the region.

Panel	Description
Running DS	The number of concurrent running DS per fleet and state
Running DS (Stacked View)	The number of concurrent running DS per fleet and the status displayed in a stack view
Namespace claim attempts rate	This is number of claim attempts on a regional basis. Measured from the AMS Fleet Commander Service.
Namespace claim success rate	This is the number of successful claims attempts on a regional basis. Measured from the AMS Fleet Commander Service.
Namespace claim failure rate	This is the number of failed claim attempts. Measured from the AMS Fleet Commander Service. Generally, a small number of errors are okay as session claims are retried. If this number is increasing or going up consider increasing the buffer of your fleet in this region.