What preparations should I make before integration?
Does the Region parameter in CAR TencentCloud APIs indicate the concurrency pack region?
No.
Region is a common parameter of CAR TencentCloud APIs and doesn't need to be specified, as CAR will select the optimal access region based on UserIP.Does CAR support a user queue?
What do UserId and RequestId mean respectively?
UserId is the custom unique user identifier passed in to CAR, such as user123456. After CAR receives the requested TencentCloud API, it will return a RequestId, such as 01fdc815-c4e7-4642-819e-a011856dfd5a1.How do I view RequestId?
If you are using a test or quick launch link, you can open the debug panel (
Ctrl+～) in the toolbar displayed on the application's main page and view the RequestId information at the bottom.If you use Chrome DevTools, you can get the
RequestId of CreateSession on the Network tab.If you have integrated the CAR SDK, the returned values of the TencentCloud API include
RequestId. We recommend you record it on the business backend.How do I view network, CPU/GPU usage, etc.? How to select an appropriate concurrency specification?
You can open the debug panel (
Ctrl+～) in the toolbar on the Test or Quick Launch page. You should pay attention to the following:FPS (frame rate): In normal cases, the frame rate should be kept above 30. If you find that the frame rate suddenly drops when you enter certain scenes or perform certain operations, it may be due to a sudden increase in GPU computing power consumptionthat exceeds the capacity of the current concurrency specification and results in frame drop. In this case, we recommend you try using a higher concurrency specification.
RTT (round-trip time): When the RTT value exceeds 100 ms, it may result in noticeable latency. We recommend you first check for local network jitter. You can try accessing over a 4G/5G network. Additionally, if the concurrency region is too far from your physical location, it can also result in high RTT.
Region: The region where the concurrency is located. CAR can automatically schedule the nearest available concurrency based on the IP address of the end user. If you find that the concurrency region is too far away, it may be due to insufficient number of available concurrencies that are closer to the end user. You can contact us for assistance.
InstanceType (concurrency specification): CAR supports four concurrency specifications: S, M, L, and XL, which are suitable for rendering small, medium-sized, large, and extra large applications. For details, see the official billing documentation. If your application requires high computing power but is using a lower concurrency specification, it may lead to high CPU/GPU usage, causing lags or forced quits.
CpuUsage: If you notice a drop in FPS data, check if there is a CPU usage of 90-100%. If so, the current concurrency specification is insufficient. Try using a higher concurrency specification.
GpuUsage: Pay attention to the value of L (load). If you notice a drop in FPS data, check if there is an L value of 90-100%. If so, the current concurrency specification is insufficient. Try using a higher concurrency specification.
RequestId: If you have any questions, try to persist in connecting and contact us with the
RequestId.Can I get the public IP of the CAR resources to implement access allowlists or other features?
No. The public IP provided by CAR is not fixed, so you need to open UDP port 8000 to all source IPs. If there are no special security problems, we recommend you open all UDP ports.
Yes
No
Was this page helpful?