In the last couple of years, many businesses have switched out to using more and more cloud services. For my company this came down to pretty much all infrastructure getting moved to the cloud. People check e-mails via office.com, have meetings via zoom and developers VPN into the virtual appliance running in AWS to access their test environments.
This means that when we get back to office, the network there has to be basically an internet cafe. But the business does rely on that internet link for pretty much everything. Yes, I have dual internet connections in offices, but right now the only monitoring that's happening is rpm probes that ping some common IPs and if there are too many failures - internet link gets switched over to another one. This is hardly sufficient for todays world.
So my question is - how do you guys monitor internet connection to all the cloud services and make routing decisions based on that? For example, I could use something like Thousand Eyes, have 2 instances pinned to different internet links and monitor connectivity through it via all of their built in tests that can cover all the cloud services I care about. I could write a script that would trigger route failover based on the tests from 1k eyes, but as far as I am aware I'd need to be able to reach 1k eyes web site to get the data that link is down - which is kind of hard to do when your internet is down. I guess I could do it from a known IP on the internet directly to the working WAN IP, but that seems a bit like a hack.
Are there other products that can do similar things? Or maybe even open source projects? I am also open to looking at some other vendors for WAN connectivity, not that I mind the Juniper SRX that I have now, but making internet failover and monitoring as easy as possible would save me from lots of headaches...
No comments:
Post a Comment