Adeptia application settings can be optimized to ensure maximum performance. Some key settings are:
Kernel JVM
- All data flows are run in the Kernel component so its JVM Heap, GC etc. are key for using available resources.
Queue Processor
- This ensures reliability by preventing spike of concurrent jobs from crashing the system. Optimize to prevent over-throttling.
Log Cleanup and Repository Archival
- Millions of entries in log tables, tens of thousands of old files in Repository can slow the overall performance.
Others details and settings of various scenarios are listed in the Performance Tuning Guide
1. Java Heap Size
Java Heap Management
- Typically both WebRunner and Kernel run in their own JVM in a single VM/server
- In idle state, due to running events, Kernel memory may go to 2-3GB
- Recommend ratio of total system RAM - 60% for Kernel java heap, 20% for WebRunner, 20% for OS
- For 8GB RAM, you may configure Kernel - 4.8GB and WebRunner - 1.6GB
- For 16GB RAM, you may configure Kernel - 10GB and WebRunner - 3.2GB
2. Garbage Control Settings
- Adeptia recommends using Oracle guidelines for JRE GC tuning for aggressive GC invocation for high memory large volume transaction processing
- Additionally Adeptia recommends using -XX:-UseGCOverheadLimitsetting to bypass GC related process flow termination
3. Queue Processor
- It prevents too many flows from being executed at the same time, overwhelming the system
- Triggered flows are automatically queued and Queue Processor releases flows to be executed to ensure only a specified max number of flows are run concurrently
- Static Queue Processor - setting of Concurrent Flows determines how many jobs are issued
- Dynamic Queue Processor – resource based allocation issues jobs if resources are available
- The key settings needed to provide this check are as below
abpm.node.resource.cpu.free **set this to to numeric for checking % of CPU availability needed**
abpm.node.resource.memory.free**Provide a numeric value in MB threshold needed for resource availability check**
abpm.node.resource.repositorydiskspace.free **Provide a numeric value in MB threshold needed for disk space availability check**
abpm.node.resourceavailability.check.waitingperiod**Provide a numeric value in secs for waiting period of resource availability check**
- PF Pause - feature to pause PF execution if an unexpected large file or data set is encountered that would exceed RAM and CPU thresholds. This ensures processes do not fail for lack of resource availability.
4. Other Parameters
Guide | Description |
---|---|
Recommended software and hardware configuration of Adeptia Suite for better performance. | |
Configuration settings to use Adeptia Suite according to your Adeptia Suite Edition. | |
Recommended settings for Java heap size for Kernel and WebRunner. | |
Recommended configurations for log and backend database connection pool. | |
Configuring logging levels for higher performance. | |
Configuring process flow for higher performance. | |
Configurations for handling a large volume data. | |
Configurations for handling the large number of transactions. | |
Increasing jetty count for handling a large numberer of concurrent users. | |
Adjusting the value of two parameters to avoid delay of the scheduled events. | |
Tuning the performance of Adeptia Suite for maximum output. |