movEAX/mondrian.properties optimal

## mondrian.properties optimal

    mondrian.rolap.queryTimeout - we set it typically to 60 seconds to cancel too long running ad-hoc queries which generate too much load on the server
    mondrian.rolap.maxQueryThreads - increase it to higher number to allow more parallel Mondrian query threads if you have many concurrent users
    mondrian.rolap.cellBatchSize - we have been experimenting with this value but typically set it to 100000 (which is now the default value for Mondrian), by changing it you can modify batch size when retrieving large set of result cells
    mondrian.native.ExpandNonNative - set to true to expand more MDX expressions to SQL
    mondrian.rolap.evaluate.MaxEvalDepth - we set it to 20 as otherwise sometimes more complex MDX calculated members were failing
    mondrian.rolap.EnableInMemoryRollup - we set it to false as we have some multi-valued dimensions which cannot be rolled up by making sum of measure values for detailed members therefore we always want rollup to be calculated using SQL
    mondrian.util.memoryMonitor.enable - we have experimented with enabling it to avoid too large Java heap memory consumption by Mondrian (if after Java garbage collection less than 10% free memory will remain then Mondrian will cancel all query threads).

	mondrian.rolap.queryTimeout - we set it typically to 60 seconds to cancel too long running ad-hoc queries which generate too much load on the server
	mondrian.rolap.maxQueryThreads - increase it to higher number to allow more parallel Mondrian query threads if you have many concurrent users
	mondrian.rolap.cellBatchSize - we have been experimenting with this value but typically set it to 100000 (which is now the default value for Mondrian), by changing it you can modify batch size when retrieving large set of result cells
	mondrian.native.ExpandNonNative - set to true to expand more MDX expressions to SQL
	mondrian.rolap.evaluate.MaxEvalDepth - we set it to 20 as otherwise sometimes more complex MDX calculated members were failing
	mondrian.rolap.EnableInMemoryRollup - we set it to false as we have some multi-valued dimensions which cannot be rolled up by making sum of measure values for detailed members therefore we always want rollup to be calculated using SQL
	mondrian.util.memoryMonitor.enable - we have experimented with enabling it to avoid too large Java heap memory consumption by Mondrian (if after Java garbage collection less than 10% free memory will remain then Mondrian will cancel all query threads).