Cameyo Analytics is a reporting and analysis feature allowing you to gather statistical data about your servers, applications, and sessions, designed for optimizing and lowering hosting costs.
Note: the Analytics feature is only available to Self-Hosted and BYO Cloud Cameyo customers. Cameyo Fully-Hosted customers are managed by Cameyo's technical team who is responsible of server optimization and maintenance, letting you focus on what's important: your applications.
Analytics Homepage
On the analytics homepage you get to see overall charts regarding your aggregated account's data, followed by tables that list the data in a sortable way:
Time range
At the top right of the screen, you can select the period to be analyzed:
For cost savings and usage pattern analysis it is recommended to select the last 7 days.
For technical support and user issue coorelation, it is recommended to pick the last few hours.
For monthly reporting and general overview, you can select up to 30 days.
To select a custom period, select the time range using the mouse. The period will the be set accordingly for the entire page:
Drilling down
From the analytics home page, you can drill down into specific servers, clusters, users and applications to obtain a more detailed view on each item individually. For this, simply click on the link you'd like to explore within the relevant list:
Cost optimization
Whether you are managing on-premises or cloud infrastructure, a cost is associated with the computing power you provide. Cameyo Analytics helps you optimize the computing power thereby reducing costs. When working on cost optimization, the recommended time period is 7 days, as it usually encapsulates a typical week.
The CPU/RAM usage chart provides actionable data to help you reduce costs, and the first one you should look into:
Below are some of the most common optimizable scenarios, along with recommended actions.
Overpower
Meaning: servers with low CPU/RAM usage.
Recommendation: either reduce the CPU/RAM amount, or reduce the cluster's number of servers.
Example: in the below example, server is in the Overpower scenario. As a first optimization iteration, CPU and RAM were lowered on May 16th, thereby increasing capacity efficiency to 55%. Additional optimization iterations can still be performed to further optimize this cluster:
Imbalance
Meaning: servers with disparity between CPU and RAM usage (CPU usage too low compared to RAM or vice versa).
Recommendation: either reduce the over-powered resource (CPU or RAM), or increase the over-used resource.
Example: in the below example, server is in the Imbalance scenario. In this example the CPU can be reduced:
Underpower
Meaning: servers with too high CPU/RAM usage.
Recommendation: increase CPU/RAM amount, or add more servers to the cluster.
Example: not an cost optimization scenario, Underpower often predicts technical issues as servers are close to reaching their maximum capacity. Server should not remain above 80% capacity -- either CPU or RAM -- at the risk of not being able to operate properly and cause usability, slowness or connectivity issues. Below are examples of underpowered servers.
Optimization deployment plan
Optimization of a production system should be applied in small iterations rather than all-at-once changes. For example if you have an Overpower scenario around 30% and you would like to increase capacity efficiency 75%, you should aim at acting in several iterations, i.e. 3 iterations across 3 days each. For example:
- Iteration 1 - aim at 45%: implement the server capacity change -> leave running for 3 days, see that there's no issues.
- Iteration 2 - aim at 60%: implement the server capacity change -> leave running for 3 days, see that there's no issues.
- Iteration 3 - aim at 75%: implement the server capacity change -> leave running for 3 days, see that there's no issues.
Technical issue spotting
The Session requests chart shows users session requests and their outcome. Unusual failure or cancellation spikes could indicate a technical issue:
On the other hand, a healthy chart typically indicates that things are working well:
This chart is especially useful for IT and support teams:
- Proactively: a regular glance at this chart helps proactively identifying issues that users are running into.
- Reactively: when users open support tickets, this chart helps understanding whether the reported issue is local / user-specific or global.