We have a reporting platform that we're trying to leverage the solarwinds API to query data from. It uses a homegrown tool that can use an API to ingest data. This works great for availability, cpu, disk, memory and response time statistics.
However, we're running into issues with interface related tables where the queries are timing out and the SWIS service itself starts to consume over 12GB of ram and then doesn't go back down until it is restarted, here is an example of one query:
select Nodes.Caption as DEVICE_NAME, Nodes.IP_Address as DEVICE_IP, 'InPercentUtil' as PARAMETER_NAME, InterfaceTraffic.DateTime as CHECK_TIME
,InterfaceTraffic.InPercentUtil as OUTPUT_DATA, interfaces.TypeDescription as
OUTPUT_DATA2, Interfaces.Caption as OUTPUT_DATA3
from orion.Nodes, orion.npm.InterfaceTraffic, orion.npm.Interfaces
where InterfaceTraffic.nodeID=Nodes.NodeID AND interfaces.nodeid=nodes.nodeid and MINUTEDIFF(InterfaceTraffic.DateTime,getdate()) = 1
tdanner, would you have any suggestions?
Id say we have a medium to large deployment and interfaces are definitely the biggest portion of the environment:
3 Polling Engines
Network Elements 32814