September was our highest load in the entire year so far, ~37% above our previous record high point this year. Its interesting that this seems to be across almost all branches and even exceeds the load during the last few months of the FF3.5 development cycle?!?
.
This wildly exceeds our expected load. Personally, I’m impressed our systems stayed up and working correctly, even if they sometimes got backlogged. Last month’s data has us trying to figure out answers to two questions:
- What was so special about September to cause this?
- What will October look like?
easy.
TM-MC, MC-TM, TM-MC, MC-TM, TM-MC, Branched, lots of regression fixes, New Features, a lot more Try-build pushes, testing mega patches, multipatch bugs, tons of code cleanups, code rewrites. And just hoping we get some good releases out the door by having more bugs filed/tested and eyes working everyday.
Basically changes to almost every feature on the trunk.
Is there a way that developers could keep track of freeze times and red burning times?
I believe it will be interesting to see how much “redness” and average freeze time has changed in the last months
Everyone – and yes, I mean everyone is working towards new releases now: Firefox 3.6, Fennec 1.0 (or whatever it will be called in the end), Thunderbird 3.0, SeaMonkey 2.0 are those I see that directly affect those trees.
This means that a lot of smaller fixes are coming in to polish up the releases, and some people do extra effort to get long-going patches landed now so they can make those releases. In the case of Thunderbird and SeaMonkey this has meant a couple of Mozilla platform fixes we wanted in 1.9.1.4, which means they needed to land on mozilla-central, then on 1.9.2 and then on 1.9.1. There was more effort being put from other sides into getting things into 1.9.1.4 for other reasons as well, it’s quite large for a security release overall – I guess that’s what we’re seeing with slightly higher 1.9.1 numbers there. In parallel to that, 1.9.2 is really coming to life, showing the work towards Firefox 3.6 and the mobile release – the latter probably accounts for mobile-browser rising as well.
mozilla-central is roughly where it always is, compare e.g. to the first 3-4 months of the year. Nothing very abnormal on places or electrolysis from what I see, the former is going more silent, the latter is steadily going on.
The releases and some calls to more discipline are probably also what makes try have lots of jobs to do, which just shows that people are doing lots of development and being careful with first checking for try results before landing more risky patches.
All in all, the overall sharp rise seems to be a quite natural result of the trees we have, edged up somewhat due to everyone trying to make releases in October and November.
What this means for October is pretty unclear – I’d expect 1.9.1 to be more calm, but the others probably still will be hit by people eager to make releases…