Finding the culprit for TCP_StatsCollector ConnectionsEstablished alert for Windows collectors
From the collector’s device page in the LM Portal or the collectors page, get to a debug console, then here’s your !POSH one-liner to get info about the destination device that is holding your ports captive. netstat -an| sls establish | foreach { ($_ -split "\s+")[3] } | group | sort count | select count, name -last 10 In the Netstat, a shows all, n shows IP addresses rather than solving the DNS for it. TheSelect-String (aliased as sls)passes only the “Established” connection entries from the netstat down the pipeline. The foreach{} splits each line ($_ is the current object being iterated by the foreach loop) on contiguous whitespace (I use this a lot!) and takes the third element (remote address:port) to passdown the pipeline It then passes Group-Object (aliased as group) which bundles identical strings and Sort-Object (aliased as sort)by the count property of the group object. The select displays grabs the calculated match count and the name properties to limit display and just shows the -last 10 of them (which are the biggest number of matched lines due to the sort previously applied. This should give you the target/s for troubleshooting further.
How many collectors do you have?
Hello, Our Logic Monitor environment is overloaded. What is the best practices to expand? We have currenty 64 collectors “All the same size” and ourinternal LM Support for the collectors is requesting 52 new collectors!! They are taking the approach to create collectors all the same size for 10,000 instances. I was thinking about other routes 1. bigger collectors if possible for 30,000 instances each may be... 2. minimize the number of instances by cleaning up what is not needed but I am not sure anybody will know We have an AZURE Collector VZPOHIALGCMON01 with 82,305 instances !!! I reviewed this Collector Capacity | LogicMonitorand apparenlty we could expand another way than the one selected !!! Any thoughts on this? Thanks, DomUse AWS Unified Cloudwatch Agent to replace Logic Monitor Collector in multiple AWS accounts
Is there any way to use the AWS Unified CloudWatch Agent and Cloudwatch Logs load EC2 metrics into LogicMonitor similar to what is pulled from theLogicMonitor collector? A little background. We currently use two LogicMonitor collectors to gather Windows metrics from AWS EC2 instances and physical and virtual servers in our on-prem datacenter. This has worked well for the 100 Windows servers and network devices we are monitoring. We are moving all of our on-prem Windows applications to AWS, which requires we change the way we use AWS. Take a look at AWS Control Tower for more details Instead of one AWS account and one VPC ( network ) , we are moving every application and environment ( PROD, UAT, DEV, STAGE ) into different AWS account. No shared networking, so VPC peering or using a TGW isn't an option. This will result in about 20 new AWS accounts and up to 100 by the end of the year. Each of these accounts will have 2 to 20 EC2 Windows instances for web/application along with one EC2 or RDS based Microsoft SQL Server instance. I have no desire to pay for EC2 Windows instances for every account ( possibly hundreds ) just to run the LogicMonitor collector. The most important Windows server monitoring features we rely on in LogicMonitor boil down to - server is down ( ping loss ) - low disk space - high CPU utilization - the ability to alert on Windows Event log entries - Windows service X is not running Then tack on monitoring of an EC2 and RDS Microsoft SQL Server performance data Does LogicMonitor have any recommendations or examples on this type of configuration? Thanks in advance, Kevin Foley
