Monitor the OpsProb mailing list for problem reports and make sure they are getting taken care of.
Attend the morning shift meeting and take note of any problems and/or report on the status of any outstanding problems.
Be available (via phone or page) to respond to any urgent problems. In principle attempt to fix these problems using the list of known problems or the how to fix pages. If problem cannot be immediately solved contact the appropriate expert.
We will continue to refine the role of the infrastructure shifts as we gain more experience, suggestions are welcome.
Since we have had little need to use the how to fix documentation up to now it is still quite rough. Please feel free to suggest improvements (or even better if you find things which are unclear try to update the documentation to make it clearer)
Some other mailing lists you might consider joining