Wherobots Usage Dashboard Guide¶
Overview: What is the Wherobots Usage Dashboard?¶
The Wherobots Usage Dashboard provides developers with a clear view of quotas, Notebooks, Jobs, and resource allocation into a single interface.
With the dashboard, you have a hands-on approach to optimizing and monitoring your system's health and efficiency.
Access & Navigation¶
- Reaching the Dashboard
- Go to Wherobots Usage Dashboard.
- Selecting a Time Frame
- Choose the desired time period for analysis.
Note
The current version of the dashboard displays data up to a maximum of 30 days old. However, this constraint is temporary and will be addressed in upcoming updates.
Metrics Breakdown¶
Please refer to organisation info to learn more about the quotas.
CPU Cores Reserved¶
Description: This graph portrays the quantity of CPU cores that have been set aside or 'reserved' over a given time frame.
Importance: Monitoring the CPU cores is essential as it directly correlates with the system's capacity for parallel processing. In distributed computing scenarios, more reserved cores can lead to efficient parallel task execution.
Kubernetes Overhead: Kubernetes utilizes some reserved CPU for its operations, reducing the actual cores available for tasks. For instance, from 4 reserved cores, Kubernetes typically uses some overhead, leaving about 3.62 cores for tasks.
Jupyter Notebook Calculation Example:
CPU cores reserved = core of notebook driver + (Jupyter Kernels × executors × cores per executor)
Given:
- Notebook driver core: 1
- Jupyter Kernels: 2
- Executors per Kernel: 3
- Cores per Executor: 4
Calculation:
CPU cores reserved = 1 + (2 × 3 × 4) = 25
After deducting Kubernetes overhead: 25 × (3.62/4) = 23.05 cores available
Memory Reserved¶
Description: This graph indicates the volume (in GB) of memory reserved during a specific period.
Importance: Adequate memory allocation is key to preventing system slowdowns or task failures. Regularly reviewing this metric helps in adjusting memory allocation according to the demands of the geospatial tasks at hand.
Kubernetes Overhead: Kubernetes uses a portion of the reserved memory for its operational needs. For instance, when 16GB of memory is reserved, Kubernetes might use 2.48GB for its overhead, reducing the actual memory available to 13.52GB for tasks.
Jupyter Notebook Calculation Example:
Memory reserved = memory of notebook driver + (Jupyter Kernels × executors × memory per executor)
Given:
- Notebook driver memory: 4 GB
- Jupyter Kernels: 2
- Executors per Kernel: 3
- Memory per Executor: 4 GB
Calculation:
Memory reserved = 4 + (2 × 3 × 4) = 28 GB
After deducting Kubernetes overhead: 28 * (13.52/16) GB = 23.66 GB available for tasks
Notebooks¶
Description: Displays the most recent or the current Notebook Instance.
Jobs¶
Description: Displays the number of jobs you have submitted.
