Hi, I’m currently working for a large Healthcare provider in the US (my background is more traditional networking). Multi/hybrid cloud computing is a fact of life. Not sure if this is the right place to ask but are there any best practices (tools) for monitoring these cloud connected environments?
It seems like you are in good company, I was just reading how cloud is gaining steam in the healthcare space: https://www.healthcaredive.com/news/cnbc-amazon-cerner-close-to-closing-on-deal/511605/
Monitoring cloud network connections over the traditional Internet model can be an IT nightmare. If your users are accessing the cloud over the public Internet, then every egress point of your network becomes a point of interest, the Internet is a black box you can’t get much visibility into, and the cloud environment doesn’t have your traditional monitoring tools. When a user reports, “I can’t access the application!” How can a network team troubleshoot the issue if they lack end-to-end visibility of the network path? If I’ve learned anything in my career, it’s that until the network team can prove it’s not the network… the network gets blamed.
VPNs over the Internet give some visibility—you can now see when the tunnel is down but you’re still left to troubleshooting the Internet and the cloud half of the VPN connection doesn’t offer much in the way of visibility.
The IOA cloud connection model gives the network team a break by helping to isolate and apply visibility and control to every path to the clouds. By cutting out the Internet and using private paths to the cloud providers, network teams can apply their own well-known monitoring tools: SPANs, Netflow, SNMP, and any other open or proprietary tools can provide detailed information on cloud bound traffic right up to the point it enters the cloud network.
If you’re using the Equinix Cloud Exchange for the network connectivity, port bandwidth monitoring has recently been added to the portal so you may check current and historical bandwidth utilization to help with capacity management.
There’s opportunity to extend visibility inside the cloud as well. Cloud specific monitor tools like Cloudwatch and Stackdriver can make sense of logs coming out of the cloud and give some insight to the cloud networks. Of course, both these tools were not built for network monitoring so tuning them to report on the important aspects may take some time. Many enterprises are deploying VNFs into the cloud environment to extend their enterprise networking and monitoring directly into the cloud. For example, Cisco has their CSR1000v and a reference architecture for Transit VPCs within AWS: https://blogs.cisco.com/enterprise/scale-to-hundreds-of-vpcs-easily-with-cisco-and-the-transit-vpc. Other enterprise networking vendors have also made their software available to be deployed into AWS and other public cloud providers as well. With a well-known networking OS running in the cloud, the classic monitoring tools can be applied to it and full end-to-end visibility can be in the hands of the networking team. Now they can quickly prove it’s not the network when that application doesn’t load!
IT nightmare is correct!!! Thanks I will do some additional research on the elements you have recommended
Choose a location