On monday there will be some scheduled, planned maintenance being performed on some of the core Metrics backend. For most of you it shouldn’t cause much inconvenience, but so that you know what’s going on, and in the interests of being as open about the system as possible, let me tell you what’s going to happen and why…
What’s Going to Happen
- At 10am PDT on Monday, Metrics will go offline for 30-60mins while a file system copy of all the databases is performed.
- When that’s done, Metrics will come back online. During this time, and throughout thewhole operation, data will continue to be collected.
- The next 6-12hrs will be spent running a conversion script to move us to a new database schema (more on why later..)
- During this time, processing (turning raw data into meaningful statistics) will be stopped, and once the above task is done, it takes about 6-12hrs to fully catch up again.
- Everything should then be back to normal
Why We’re Doing This
The difference between theory, and real live data is vast. Now we have around a months worth of real data to analyze, we’ve found that we can make improvements to the system that just hadn’t been noticed in the theory part of Metric’s development.
We’d like to make those improvements for the good of the system, and hope you’ll bear with us during Monday while this work is done.
What Comes Next?
A few weeks down the road, there will be another tweak similar to this. I’m hoping we can also learn from this and fine tune the process so it has even less impact on users, but i’ll have to talk to you about that nearer the time when more info is available.