Consider the technical information about this metric subnode in the metric hierarchy:
| Software Product and Version | SAP Convergent Charging 5.0 |
| Path | SAP Convergent Charging | SAP CC | Queue |
This node groups the performance and health data about the main processing queues and their activities in the SAP CC Core Server systems. Consider the queues regarding to:
Each subnode represents a specific processing queue in an SAP CC system instance. Each queue processes specialized items using a thread pool.
| Domain | Monitored Queue (Technical Name) |
|---|---|
| Services based on Message TCP | Stateful, Batch, Cold |
| Response handling in dispatcher instances | Guider Response, Rater Response |
| Multithreaded processes (Warm-up, Activation, Allowance Purge, Cleanup) | Allowance |
For parallel processing and performance optimization, SAP Convergent Charging handles few queues and corresponding mechanisms. You cannot monitor all the queues because the SAP CC system does not report any performance data.
Important Note
The final usage of a queue depends on the instance type that manages the queue. For example, the final usage of the Batch queue differs in a rater instance and in a guider instance.
A queue is a collection of items in a specific order. The queues are used for interthread, nonblocking, asynchronous communication. Consider the queues regarding to throughput, latency, and reliability.
The items are commonly enqueued by a specialized thread, for example responsible to read sockets. The items are dequeued by processing threads commonly managed in the thread pool. The items are dequeued before their processing, meaning that if a queue is blocked, then it is not related to its content (head), it is related to the last dequeued item or items in case of batching. The queues could be size-constrained and so subject to specific errors when the queue is full.
The queues (and multicast services) do not provide "one-to-many" mechanism. The items (and services) are "one-to-one" meaning that there is only one consumer/receiver for one provider/sender.
All queues are in memory meaning that if the cluster goes down then the items are lost. Since the items could be forwarded (routed) there is some mechanism to retry or reply properly.
Some queues watch their items and process the expired one based on their timeout. The item still resides in queue, even after it has been replied with the common timeout status. See also the asynchronous stateful service client timeout setting and the round trip time parameter.
The queues offer the possibility to batch dequeue. The batch dequeue is managed by the consumer (see stateful service request batch size).
At the cluster level
The queues are not used in a "send and forget" mode meaning that the originating item
(cluster level) is kept.
Note: The generated data file (bulk write) are queued using a "send and forget" mode in
this case items are kept in case of error and file (queue) could be blocked.
The items forwarded to another queue in another instance are kept in the originating instance here the dispatcher. The dispatcher owns the client connection so is in charge of the reply. The items are not kept in the dispatcher service queues (entry point). the items are moved in other queues such as connection (gargeable) and response
The receiver may also be a sender (routing) for the same or new items. in this case this receiver provides and ensures the coherency of the item handling and failure. The receiver routes the item programmatically and based on item content.
The messaging layer offers ways to report queue full, timeout, and failures out of the receiver processing.
Note
Not all queues are monitored nor reported. Among the notable missing queues:
Example
The following is an example of the queues part of the charging operation originated
by the asynchronous stateful service client.
Note: See also the flow control available for this specific client.
Note
There is no metric defined in SAP CC at this level. Consider the subnodes in the metric hierarchy.
The possible subnodes are:
For more information, consult the following user assistance:
