ReportCaster Connections - what exactly is this number?

[Debra Waybright]

I'm looking at the Server Workspaces screen and see a column "Connections" and I'm just curious what this number actually represents. We have some reports that have 43 and a few that have 82 and I'm wondering how this relates to the maximum threads setting in the ReportCaster configuration. That number, for us, is 6. That's a lot less than 82! I'm curious what the correlation is and how we might be able to use this information to improve our system.

 

[Patrick Huebgen]

@Debra Waybright - please share a screenshot - I'm not sure which screen / setting you are referring to?

[Clayton Peacock]

Hi Debra, 

I think you are referring to the agent refresh which defines the maximum number of new connections which can be accepted during the life of each agent process - default is 100. 

Beyond this limit, additional new connections will be assigned to a fresh agent. 

Agent processes which have reached the limit will be terminated when the last accepted session disconnects.

A server configuration requires at least one agent service with the name DEFAULT, defined by a SERVICE block. An agent service is an entity used to define the parameters for a group of data access agents, so that a configuration can manage different groups of data access agents for different purposes. Each data access agent runs for a specific Data Service, and each service may have different values for the settings defined on the services configuration panes. These settings include:

The maximum number of data access agents and the number of agents prestarted at server startup for a service, as defined by maximum and number ready.

The lifetime of the agents of a server, which can be limited through idle agent limit, CPU time limit, and/or memory limit.

Incoming connections for which there is no available data access agent. These can be put in a queue for the service (configured using maximum_q and queue limit), and after they are connected, their idle time can be limited using idle session limit.

The deployment mode of a service, which defines how data access agents are assigned to connections.

Kind Regards,
Clayton 

[Debra Waybright]

@Clayton Peacock Yes - that is what I am looking at. Where is the setting that you show at the bottom of your post? 

Also, "Agent processes which have reached the limit will be terminated when the last accepted session disconnects." -- Does this mean if a process attempts to get more than the maximum agents configured will crash? 

 

Thanks!

[Clayton Peacock]

 

On the workspace screen:

• Right Click Data services
• Select Properties

Note that there is an on hover help next to each of the settings. Also, in the top right of the server console there is a (Online Server Help) select ? (question mark) icon to expose drop menu this will give you more detail.

"Does this mean if a process attempts to get more than the maximum agents configured will crash?" - No, this means that the agent will be "refreshed" start from 0 again.

My recommendation is not to change these settings.

[Debra Waybright]

I do have another related question - are there tricks for reducing the number of agents a schedule uses?  Our maximum number is 100 and we have schedules that show 83 connections when running. I assume that means another schedule that needs more than 17 agents won't be able to run until that first one is complete. If we could reduce the number the first one uses, it would allow more schedules to run at the same time (again, I'm assuming). 

[Clayton Peacock]

Hi Debra, 

One ReportCaster task should use one agent not 83. Once the agent is finished processing the task, the connection that you see in the server consoles "connections" will increment till it reaches the agent refresh limit of 100 and will recycle the agent.

You might want to increase the number of max agents available if all your agents are being utilized - maybe increase to 50 maximum but be cautious as this will take up server resources possibly.

Are you ReportCaster jobs queuing that you are concerns about this?

Maximum:
The number of agents available ready(default) is 40
Defines the maximum number of data access agents that the Workspace Manager will allow to run simultaneously for a specific service 
The absolute upper limit for the value of this keyword is the size of the agent table and may be limited by system resources. 
But in practice it should be significantly less than that to leave enough room in the agent table for agents in transitory states while allowing for the maximum number of active agents.

Number ready:
Defines the number of data access agents that the Workspace Manager will create at startup for a specific service. 
These pre-started agents will thus be available in advance so that client connections can be established more quickly, due to the preloading of some modules. 
The Workspace Manager will not attempt to maintain this number of agents after startup if agents are terminated for any reason.

Regards,
Clayton 

[Debra Waybright]

@Clayton Peacock Yes, it seems that when we have one of these schedules showing 83 connections maybe only one other job is able to run and so people are waiting for their reports. These jobs are typically ones that are checking for a condition and sleeping then checking again until the condition is true (we use this to know when a process has completed by checking a logging table) but it bogs things down when it ends up checking and sleeping many times. 

[Patrick Huebgen]

The image does not show Schedules as this is not ReportCaster it‘s the Reporting Server - the 83 does mean that this agent was running 83 jobs in the past. If you need to see ReportCaster schedules please use the ReportCaster console.