Few months ago, we had the amazing opportunity to deploy SexiGraf in a *BIG* company. Think petabytes scale.

In the past, we have already been challenged to make our tool scalable, but not that much:

  • 50+ vCenter
  • 6000+ ESX
  • 115000+ VM
  • 18000+ datastores
  • 1000+ clusters
  • 200ms+ network latency for half vCenter

After many hours of hard work, almost total rewrite of our code and a massive use of hash tables, we are proud to show you a sneak peek of the next SexiGraf version (0.99f) connected to 53 vCenters:


And since the customer hasn’t figured out how to handle syslog for it’s 6000+ ESX yet, we added a “bad” events dashboard (warnings and errors) grouped by clusters:

With a single VM (16 vCPU and 24GB vRAM), the polling is only taking 3 minutes (maximum) to complete:


Stay tuned 😉

Leave a Reply