JD Edwards AIS server tuning: timeouts, response size and orchestration performance

The AIS and HTML server settings to review when JD Edwards orchestrations, UX One or REST integrations slow down — with the Oracle Support notes behind each recommendation.

The AIS server is the REST entry point of JD Edwards EnterpriseOne: orchestrations, UX One, watchlists and every external integration go through it. These are the settings we review first when it struggles, each with the Oracle Support note it comes from. All are set in Server Manager.

Maximum REST response size (downloads)

Reference: What is the Significance of the MaxExternalResponseSize Parameter in the Rest.ini? (Doc ID 2949357.1)
Where: Configuration → General
Setting: Maximum External REST Response Size
Why: limits the size of any external REST response. Too low and large downloads fail; too high and one call can exhaust memory.

Timeouts

The rule: the AIS timeout must be greater than the timeout of the main HTML server, and the timeout of the dedicated AIS-HTML server must be greater than the AIS timeout.

AIS server

Reference: AIS Session Token Life and Timeout Settings and How They Relate to HTML Timeout Settings (Doc ID 2522434.1)
Where: AIS instance → View = Advanced → Security Information
Settings: SessionInactivityTimeoutMinutes, SessionTimeToLiveMinutes — the two timeouts that govern an AIS session.

HTML server (JAS) and dedicated AIS-HTML server

Reference: same note (Doc ID 2522434.1)
Where: HTML instance → View = Advanced → Cache
Setting: UserSession — the user session cache timeout.

Also check the session timeout of the application server itself (in WebLogic, in web.xml): E1: JAS: Best Practices on Setting User Session Timeout Value for HTML Server (JAS) Instance Configured on Any Application Server (Doc ID 1488013.1).

UX One and watchlists

Reference: Question With the Architecture of the AIS Server in Relation to the HTML Servers (Doc ID 2239069.1)
HTML instance → View = Advanced → Web Runtime: AIS Maximum Concurrent Calls — how many AIS calls are processed at the same time.
AIS instance → View = Advanced → Cache: Read Cache Time to Live (Milliseconds) — how long responses stay in cache.

Orchestration performance

Reference: Orchestration Processing Performance Recommendations (Doc ID 2587751.1)

The real load is on the HTML server: AIS is mostly a pass-through to the HTML server, where forms and data queries run. For the best scalability of orchestrations and AIS services, use a dedicated HTML server for AIS requests — or a cluster of them — and never the front-end HTML server used by interactive users.

Size the heap of both the AIS and the HTML servers for the load. In Server Manager, under Server Start → Arguments, set minimum and maximum to 4 GB on the AIS and HTML instances:

-Xms4096m -Xmx4096m