Some downtime

The service was down for a while, but is once again available. There will be some missing data until I’ve pulled the data from the secondary site and plugged the gap.

Hole-y data

So the current users graph has a few more holes than expected. I’ve updated the script to be a little more intelligent and poll when it’s not received updated statistics (when it expected them too). Linden Lab/SL statistics aren’t perfect and not updated on time and sometimes not at all, so this will hopefully fill some hole – but it won’t do them all, alas.

We’ll be changing how the statistics are processed in the near future. Currently it’s using an old feed which does not have a timestamp. This means that there isn’t a way to know when the data isn’t being retrived apart from when you get repeated data for a prolonged period of time or noticeably horizontal lines followed by vertical lines on the graphs.

This will be more accurate, but also it will mean more ‘holes’ in the graphs. We hope to change the graphing engine also as well as giving more info on problems with the data.

This might also mean posting slightly more regularly on the site. You never know…

SL is currently down to all but Lindens right now. They’re doing work on the asset servers to increase their reliability, something that everyone wants to happen.

It’s certainly a different time to pull the grid off line for a planned maintenance for Linden Lab, but it’s certainly a good thing to have these planned downtimes shifted around the day. Hopefully this will alleviate the serious problems we’ve been seeing with the systems:

Historic concurrency
Click on the graph for more detail

Stats interruption

We seem to have had some DNS issues apparently and lost about one hour, twenty minutes worth of statistic. Not sure if it was my end or Linden Lab end, but it is working again.

Well it used to be Sundays were the big problem nights but now it seems to have become Fridays. It seems that we’re told not to perform any transactions as well as logins are restricted. Concurrency numbers are holding up.. for the moment.

[Update: 8:11 GMT] Spoke to soon on the concurrency numbers..

[Update: 8:43 GMT] The blog is now reporting that logins are open again and the Asset Server is okay.

As you may have noticed I’ve changed the theme to a new one (Tarski) and tweaked it a bit for width (and purpleness). The theme is still quite simple and easy on the eyes, the pages are more obvious than on the previous theme (Pixkit) and as it’s wider I can put bigger graphs on the page.

I’ve also been working on the statistics gathering script to make it a bit more resilient when LL’s website isn’t entirely forthcoming with it’s statistics. It seems to be working happily (no holes in the graphs), though Second Life hasn’t been giving us problems like it had been recently.


I’ve moved the server to GMT as it’s a small headache for me with having to change the times. SLT is currently GMT-7, the UK GMT+1 and most of Europe is on GMT+2. Hopefully this’ll make things a bit easier.

Reduced SL Service

Well to keep SL up (more or less) Linden Lab have disabled functionality. While I can understand the need to keep SL up and running, it seems these key services (see below) are a a little too fundamental to the heart of SL with the notable exception of the first one.

I can understand disabling logins. There was some talk last year about when the grid was under heavy load that people connecting should be prioritised and I can see the logic in that far more than disabling some in world transactions. It’ll be interesting though to see how well the online user count keeps up.

Currently disabled are:

  • Logins
  • Lindex
  • In-world Transactions
  • Some of the Website

[Update 9:20pm BST] Well the concurrency graph speaks to itself. As of this moment 13,000 people have logged off or been disconnected. (This graph will refresh every 15minutes and data is added every 5 minutes)

[Update 9:40pm BST] Well the blog says that they are reenabling logins and services again. Only lost about 20,000 online users in that time…

A bad, bad day

Click for more
Click on the image above for more detail.

Well another bad day. The worst in my memory, anyway. Between routing issues (see earlier post) and in world issues the grid has been down most of the day and the night before and when available was slow and laggy. As you can see on the graph above (click for more detail) we’re down to having a closed system for Lindens only. It’s been a while since the introduced the het-grid and we’ve not really seen every being kicked off since that. Here’s hoping it’ll be fixed soon and on a more permanent basis.

[Update 9:30pm] Second Life is up once again. Watch the rate at which users log in…!