Like the title says, we arrange with our business customers for their applications and services to be unavailable for several hours over a weekend, 4 times a year. We pay so much for redundant networking gear that I feel like this is a backwards way of doing things.
One reason we do such a broad impact at one time is because we can’t seem to figure out exactly which applications/services will be impacted. Another reason is that many of our big applications are “sensitive” to losing the network. I.e. if they lose connectivity for 20s they start rebooting, or pull themselves out of the available pool for a service, etc. therefore, the support teams for those applications prefer to shut down the application for many hours just so it’s graceful, even if our impact to them might only be 30 mins.
How do you schedule disruptive work at your enterprise?
No comments:
Post a Comment