CAIOS Usage
Monitor detailed usage metrics and quotas for your CoreWeave AI Object Storage.
The CAIOS Usage dashboard provides comprehensive insights into the performance, resource utilization, and quotas of your CoreWeave AI Object Storage.
This dashboard allows you to:
- Monitor request rates and bandwidth usage.
- Track response and request durations for different object sizes.
- Evaluate storage quotas and actual usage.
Prerequisites
You must be a member of the admin, metrics, or write groups to access Grafana dashboards.
Open the dashboard
- Log in to the CoreWeave Cloud Console.
- In the left navigation, select Grafana to launch your Managed Grafana instance.
- Click Dashboards.
- Expand AI Object Storage, then choose CAIOS Usage.
- Select the appropriate filters at the top of the page.
If you're already logged in, you can directly open the CAIOS Usage dashboard from this link.
Filters and parameters
Use the following filters at the top of the dashboard to customize your data view:
| Field | Description |
|---|---|
| Data Source | Prometheus data source selector. |
| Zone | Availability zone for data viewing. |
| User | Cloud user or principal accessing the storage. |
| Bucket | Specific storage bucket to monitor. |
| Operation | Type of storage operation (e.g., read, write). |
| Response Code | HTTP response code filter. |
| Response Flags | HTTP response flags filter. |
Set the time range and refresh rate parameters at the top-right corner of the dashboard. The default time range is 24 hours.
Panel descriptions
The CAIOS Usage dashboard contains the following panels, grouped into logical sections:
Usage metrics
| Panel | Description |
|---|---|
| Total Requests per Second | Shows the rate of HTTP requests to CAIOS. |
| Bandwidth | Visualizes incoming (downloads) and outgoing (uploads) bandwidth. |
| Response Duration - Small Objects | Time between "Time to First Byte" and "Time to Last Byte" for responses ≤ 100KiB. |
| Response Duration - Small Reads | Same as above, specifically for read operations ≤ 100KiB. |
| Request Duration - Small Objects | Duration from client request initiation to completion for requests ≤ 100KiB. |
| Request Duration - Small Reads | Same as above, but for read requests ≤ 100KiB. |
The panels related to response and request durations help you monitor latency in data transfers and identify potential bottlenecks or performance issues in object storage operations.
- P50: Median duration, representing typical performance.
- P99: 99th percentile duration, indicating peak latency experienced by 1% of requests.
Regular monitoring can help optimize your storage performance and efficiently manage storage quotas and resources.
Quotas metrics
| Panel | Description |
|---|---|
| Total stored bytes quota & usage | Tracks storage quota versus actual usage in bytes, broken down by storage class and zone. |
| Bucket count quota & usage | Shows the number of buckets used versus quota allocated. |
| Stored bytes usage per bucket | Displays total stored bytes for each bucket, categorized by storage class and zone. |