Available metrics
Metrics fall into three categories:
-
Automatically collected metrics
-
SDK-provided metrics
-
Custom metrics
Automatic metrics collection
No code changes required for these metrics:
Instance metrics
| Metric Type | Description | Use Case |
|---|---|---|
| CPU | Percentage utilization per instance | Resource monitoring |
| Memory | Physical memory usage and percentage | Capacity planning |
| Network I/O | Bytes and packets sent/received | Connection health |
| Disk I/O | Read/write operations and throughput | Storage performance |
Fleet metrics
| Metric Type | Description | Use Case |
|---|---|---|
| Active Instances | Running instances count | Fleet scaling |
| Game Sessions | Active and available sessions | Capacity management |
| Crashed game sessions | Game sessions that have crashed | Error monitoring |
SDK-provided metrics
Requires SDK function calls in your code:
Server timing metrics
| Metric | Description | Implementation |
|---|---|---|
| Server Delta Time | Difference in time between the current server tick and the previous server tick. Measures the consistency of the server's tick rate | Call GetDeltaTime() |
| Server Tick Rate | Shows the number of times per second the server is processing updates | Automatically calculated |
| Server Tick Time | The amount of time it takes for the server to process a single tick or update | Call GetTickTime() |
| Server World Tick Time | The amount of time it takes for the server to update the game world with each tick | Call GetWorldUpdateTime() |
Implementation: For engine-agnostic SDKs (C++, C#, Go), you implement these metrics by calling SDK functions from your game loop with calculated timing values. For engine plugins (Unreal, Unity), these metrics are captured automatically through engine integration.
Network metrics
| Metric | Description | Implementation |
|---|---|---|
| Connections | The total number of network connections the server has established | Automatic after InitMetrics() |
| Network I/O (Bytes) | The total number of bytes being sent and received by the server over the network | Automatic after InitMetrics() |
| Network I/O (Packets) | The total number of network packets being sent and received by the server | Automatic after InitMetrics() |
| Packet Loss | The percentage of network packets that are being lost during transmission | Automatic after InitMetrics() |
Implementation: Integrate SDK function calls with your networking library. The SDK provides guidance for different network implementations.
Process metrics
| Metric | Description | Implementation |
|---|---|---|
| CPU Usage (%) | The percentage of CPU resources being utilized by the game server process | Automatic after InitMetrics() |
| Memory Usage (Units) | The total amount of memory being consumed by the server processes | Automatic after InitMetrics() |
| Physical Memory Usage (%) | The percentage of the server's total physical memory that is currently being utilized | Automatic after InitMetrics() |
| Server Status | Game server health state | Automatic after InitMetrics() |
Implementation: These metrics are automatically collected by the SDK for each game session process.
Per-process dashboard organization
Per-process metrics are available in two specialized dashboards:
-
Server Performance dashboard — Contains server timings (delta time, tick rate, tick time, world tick time), network metrics (connections, I/O bytes/packets, packet loss), memory usage, and CPU usage for individual game sessions.
-
Instance Performance dashboard — Features "Top N Memory Consuming Game Sessions" and "Top N CPU Consuming Game Sessions" tables that help identify which processes contribute most to instance resource consumption. Clicking on Game Session links enables deeper investigation of detailed metrics.
Per-process metrics use cases
The per-process/per-game-session metrics support the following monitoring scenarios:
-
Dive deep performance investigation — When a host/instance has degraded performance due to specific processes or game sessions, per-process metrics help identify which process caused the issue through Top CPU and Memory consuming Game Sessions tables.
-
Game server crash investigation — When a game session crashes, these metrics help determine if the crash was due to out of memory, CPU overload, or network bandwidth problems.
-
Investigate player reported issues — When players report lag or interruptions during gameplay, per-process metrics help identify bottlenecks in CPU, memory, network, tick time, or world update time.
-
Identify performance changes in different builds — Tick time, tick rate, and world update time metrics allow developers to measure how game performance changes across different server builds.
-
Detect delays and slowness in gameplay — Tick time, tick rate, and world update time metrics reflect how fast the server updates the game, directly impacting customer experience.
-
Benchmarking — Identify how different game scenarios affect server performance based on factors like player count, game mode, and other variables.
Dashboard organization
Metrics are organized into specialized dashboards in Amazon Managed Grafana for different monitoring scenarios. The available dashboards depend on your fleet type:
EC2 Fleet dashboards
-
EC2 Fleet Overview dashboard — High-level fleet capacity, scaling insights, concurrent players (CCU), instances, player capacity, and crashed game sessions.
-
Instances Overview dashboard — Aggregated host-level metrics across all instances including average CPU, memory, network, and disk utilization.
-
Instance Performance dashboard — Detailed metrics for individual instances with "Top N Memory Consuming Game Sessions" and "Top N CPU Consuming Game Sessions" tables for identifying resource-intensive processes.
-
Server Performance dashboard (EC2) — Game loop timing, network performance, memory, and CPU metrics for individual game sessions on EC2 instances.
Container Fleet dashboards
-
Container Fleet Overview dashboard — High-level overview of container fleet resource utilization including CPU reservation, memory utilization, and container group status.
-
Container Performance dashboard — Detailed metrics for individual containers within specific ECS tasks including CPU utilization, memory usage, network I/O, and storage performance.
-
Server Performance dashboard (Container) — Game loop timing, network performance, memory, and CPU metrics for individual game sessions in containers.
For detailed dashboard information and usage instructions, see Dashboard organization and usage.
