Custom Metric Beacon

Custom Metric Beacon allows every instance of ShowSlow to configure metrics it wants to record in addition to YSlow and PageSpeed ranking.

Once metric beacon is configured, simple GET/POST requests can be sent to the beacon with some numeric value which gets recorded and displayed over time on a graph.

You can also click on “Custom Metrics Beacon” on beacon list page (e.g. for public instance) to see interactive form for manual submission.

Custom Metrics can vary depending on your needs, good examples include:

  • bounce rate on the page (from your analytics software)
  • amount of results on search page (from system configuration)
  • server load (from backend monitoring tool)

All these metrics are specific to your needs and you have to write some tools to collect them, but ShowSlow will be able to aggregate them all along with the rest of the data.

For configuration docs, see Installation page on the wiki.

Google Page Speed talk at NYC-GTUG and Web Performance Group

NYC Google Technology Users Group
Last week I gave a talk at joint NYC-
and NY Web Performance Group event at Sun Microsystems offices!

It was very nice to see so many people interested in Google’s new performance initiative, I think many were excited and couldn’t wait to try speeding up their sites.

I posted my presentation to TechPresentations here:

and here’s direct link to the slides:

I’m trying to improve my presentation skills so please give me some feedback and rate my presentation on SpekerRate:

New York Web Performance Group Logo
A few major links from the presentation to get you started on Google Page Speed:

If you’re on Twitter and still not following me @sergeyche or the group @nywebperf, feel free to do so.

Automating Page Speed and YSlow monitoring

ShowSlow logo
A few people on the list asked me how to automate Show Slow monitoring and dug up all the magic I used on my monitoring farm and try to summarize it here:

Here are a few steps you’d want to follow:
Set up separate browser profiles – one for Google Page Speed and one for YSlow.
To learn how to set up a profile, take a look at Mozilla’s documentation: Managing profiles

Install extensions in respective profiles
Install Firebug for both profiles and Page Speed and YSlow in their respective profiles (extensions are installed and configured separately for each profile).

Set default Firebug tab
In each profile, set up it’s particular extension to be a default Firebug tab.

To do that, you need to set extensions.firebug.defaultPanelName property to:

  • pagespeed for Page Speed
  • yslow for YSlow

Configure autotesting
Configure to run measurements automatically when browser starts

  • extensions.PageSpeed.beacon.minimal.autorun = true for Page Speed
  • extensions.yslow.autorun = true for YSlow

Run Firefox periodically
Next step is to create a list of URLs you’d like to test on a regular basis and run Firefox periodically.

You can use a simple script I bundle with Show Slow:

If you’ll have any troubles, please feel free to post your questions to the discussion group and I’ll be happy to answer them:

I’m planning to add better URL management to Show Slow itself so your automation scripts can fetch a list of URLs to process right from your Show Slow instance and some other tools can post URLs to be monitored.

There is probably more to be done to simplify this automation – from better shell scripts to additional Firefox extensions. More stuff will come along eventually.

Extended Google Page Speed support in Show Slow

ShowSlow logo
I decided to extend Google Page Speed support in Show Slow to keep track of metrics the same way it’s done for YSlow.

You can see results at – here’s example with Page Speed results only:

And here’s the example for both Page Speed and YSlow together:

Version 0.3 is released and can be found in SVN here:

Please upgrade and send your comments and questions to the support group here:

Path of the most embarrassment

I’m working on HowDoable for a few months already and it’s been a fun ride and main challenge is to go towards the goal.

I’m using Pivotal tracker for my agile needs and i have a long backlog which I’m adding to more stuff over time, rearranging the list and so on. What I noticed that rearranging gets a pattern during hard times – I tend to avoid some kind of tasks – those that are not very natural for me, like writing some texts in English or just regular startup problems like writing business plan or think about the product from user’s perspective.

What I tend to push up the list is something I’m comfortable doing, like coding new features or playing around with new colors, logo, some other things, but definitely not those that bring me closer to the release.

Many words are said about shipping the product early and why it can be vital to business so I’m not going to repeat them here, but I never found a good recommendation about how exactly do we do that. One thing that I heard lately is a paradigm of “minimal viable product” which is about the set of features to include that will make the product stand out, but no more.

What it doesn’t describe is how do you do that in agile environment when so many backlog items come up when you start and reprioritization happens very often and requires constant attention. So I struggle to find that feeling that I can catch and follow during the periodic reordering of my task list / backlog to make sure that things that need to be done to be successful don’t get pushed down by things that I just feel like doing.

Some time ago I read a good quote from Reid Hoffman of LinkedIn – “If you’re not embarrassed by your first release, then you launched too late.” and it stuck with me as a great description of the problem of over-polishing the product which is not even seen by users yet.

So I though to myself – if I look at it from agile perspective and treat each iteration’s or even day’s product as a final product (thing that you want to have in your agile world) then embarrassment is exactly the feeling that needs to guide you when you’re picking the next thing to do.

What I mean is that if you thinking of moving some task down to replace it with some other task, think if this other tasks’ completion is going to help you avoid some potential embarrassment and if so, then stop right there! – what you’re doing is trying to feel better about the product and not bringing the product to the state that will benefit the business.

Unfortunately these are two different things and extreme perfectionism tha arises from trying to “polish the torpedo” never did any good and many projects and businesses were ruined by “unfinished torpedoes with shiny casing”.

So now what I do and urge all the strartupers to do (or open source developers for that matter) is to follow the path of the most embarrassment – when you feel that you have a choice of doing something to reduce the embarrassment, turn around and go into the opposite direction because you’re trying not to hurt your feelings. What you need to do instead is to go towards the business goals that I hope you had time to define and didn’t just jump into development, right.

Hope this insight helps and if you feel I’m wrong, let me know in the comments.