Networking
Settings for how the different services will communicate and interoperate.
Load balancing model
In a multi-server deployment (cluster), the internal Routing Service will automatically send requests to the various engines to share the load. Changing the load balancing settings in this panel will determine the method for sharing:
- Round Robin - Going down the list of servers in the group, the round-robin load balancer forwards a task to each server in turn. When it reaches the end of the list, the load balancer loops back and goes down the list again
- Performance Routing - The load balancer will share out tasks based on a evaluation of the load a given server is under in comparison to other servers in the cluster. This is affected both by its hardware settings (virtual or bare metal) and by the current list of tasks it is processing.
Note: This feature is only available in the Enterprise edition
Internal network encryption
Internal network encryption offers an additional layer of security, by enabling the encryption of communication between the Pyramid services internally. By default, internal network encryption is not enabled.
Note: Enabling internal encryption may affect performance
Note: This feature does NOT need to be enabled to use external encryption via SSL for the web client.
Max ping wait timeout(sec)
Indicates how long (in seconds) to wait for a service response before going to the reset flow,
Request Fail Over
Indicates what action should be performed when a run time request fails during processing.
When a request is redirected, it is sent to a different run-time engine. The options are:
- Never Redirect - In the event of a failure, no redirect will occur. The user will receive an exception.
- Always Redirect - A request will always be redirected when the processing fails.
- Selective Redirect - Queries that are using more than a set amount of CPU resources will fail.
Task Load Balancing Model
Specifies the model to be used for Task Balancing:
- Eager: This means that the task engine would take on as much of the tasks as it could handle. This would result in you seeing one task engine working really hard and a second one doing nothing.
- Semi-balanced: This means that the task engine will take on half the jobs is is capable of doing at any one time, with the idea that the second task engine will then taken on the rest. Therefore this helps spread the load better than the eager option.
Selective Threshold
When Selective Redirect is chosen for the Fail over threshold, users can set the level of CPU resources to use before a fail over occurs.
Log CPU Performance
Logs the CPU performance so that it can be analysed.
Warning: This option should only be used when absolutely needed. It will cause performance degradation when it is enabled.
Active Memory Management
- Proactive Garbage Collection: automatically free up memory by removing unused items. Every half-hour, Pyramid will check if there are unused items; if so the garbage collection will run. The garbage collector can also be run for a specified service from the Summaries Panel, or for all services from Server Instances.
- Proactive Renderer Collection: automatically clear printing processes every half-hour to free up printing services.
Proxy Settings
Many organizations use Proxy Servers as part of their internet gateway topology. Since there are various items within Pyramid that require access to the internet, the proxy server configuration may be needed to enable these functions.
Note: This feature currently supports printing of map visualizations, and querying Google BigQuery and Google Analytics data sources.
To configure a proxy server:
- Check 'Enable proxy server'
- Provide the proxy server's address and port.
- If your proxy server requires login credentials, select 'Use Credentials' and enter the server's user name and password.
- Choose whether or not to bypass the proxy server for local addresses.