How Can Large HTTP Sessions Be Identified In The Coherence*Web Session Management For An Application Server, WebLogic? (Doc ID 1980732.1)

Last updated on NOVEMBER 21, 2023

Applies to:

Goal

How Can Large HTTP Sessions Be Identified In The Coherence*Web Session Management For An Application Server, WebLogic?

There are customers reporting performance issues in the Coherence*Web managing the HTTP sessions data. Customers are interested to know how the large objects of the session are persisted, get identified and get queried from the coherence caches.

The session size per request depends on the amount of session information stored in the session. Out-Of-The-Box there is a clustered cache used to store the "overflowing" (split-out due to size) session attributes. This cache will hold large session attributes. By default, session attributes larger than 1 K will be stored in this cache. It is named as "session-overflow": This cache is used to manage the session attributes that are large enough to warrant being "split out" from the core session data, session-storage is the cache name OOTB for storing core session data.

Overflow cache, "session-overflow", is only used for the "Split" model. Suppose http session (attribute) sizes are becoming larger in the course of time, you will see the session overflow count go up as Coherence will store large attributes in it for optimized performance. You can see the *size* and *unit* attributes for the session-overflow in the JMX Mbean console. There is a coherence*web context parameter controls the use of the session-overflow cache for large objects, it is "coherence-attribute-overflow-threshold" - this value specifies the minimum length (in bytes) that the serialized form of an attribute value must be for it to be stored in the separate overflow cache that is reserved for large attributes. If unspecified, this parameter defaults to 1024.

Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.