Question on 50,000 limit

[Philipp]

This limit is true even when the chart doesn't change, right?

So when I use this service to include a pie chart in my blog post, and
the blog post gets dugg sending 60,000 visitors, then the latest
10,000 of them will see some blank image or what? (What will they see,
anyway... a file not found or...?)

Or is this limit excluding images that can be cached as their values
don't change...?

[barryhunter]

Yes I suppose this can be clarified into: is the 50k, per

1) client IP (regardless of which site they get the chart from)
2) webpage url
3) website domain
4) individual chart generated [but this could cross with other sites
who happen to create the same chart?]

??

[Herm]

The usage policy says "50,000 queries per user per day"
http://code.google.com/apis/chart/#usage
This means one dig visitor can refresh your page 50,000 times.

Regarding caching, this is done at the client side and I would assume
it would make sense that if the query string remains the same in your
<img> tag, then the client could use cache (or a transparent proxy
could use cache) instead of requesting it again from the google host.

[Hernán Rodriguez Colmeiro]

Buth that term doesn't define "user". A "user" may be a visitor to
your webpage, but your webpage URL might also be the "user". The
policy is confusing, when it should be plain and clear :(

Hernán

[uwe.m...@google.com]

The "user" is the webpage that links to the chart. We would like to
make sure that all users get their charts fast and reliably. Therefore
we may temporarily block users (websites) that exceed the limit.
A chart request is not counted if the chart image is cached by the
browser or by a proxy.

Uwe

On Dec 7, 5:25 am, "Hernán Rodriguez Colmeiro" <colme...@gmail.com>
wrote:

[wlai]

Thanks for the clarification Herm and Uwe. Does Goggle have any
intention to raise this limit over time?

[Philipp]

So, if you get 60,000 people from Digg, your chart will break, meaning
you should not include the chart in a blog post as it may get dugg?

[ao5...@wayne.edu]

If it's a blog post, and the data doesn't change, why not just query
once, then save the image to your own server? It seems kinda weird to
query Google to make the chart every time a visitor comes to your
site, considering it's actually static content.

[underbl...@gmail.com]

Well, I kinda doubt google is re-rendering these graphs for every
request. They are entirely stateless so all it should cost them is
bandwidth.

[Andrew Chilton]

Yep, each graph has an 'Expires' header which is access time plus 24
hours and 'Cache-Control' max-age parameter of 86400 which is also 24
hours.

Certainly for blog posts, I'm going to create the graph, save it and
serve it from an alternative location. I guess no-one has to do that
but it's just another option in your arsenal.

Cheers,
Andy

--
contact: Andrew Chilton
website: http://kapiti.geek.nz/

[Sudhir]

Google's policy (far as I've seen, including the Google Maps Geocoding
API) is that its 50,000 requests per CLIENT IP... so that means that
even if your blog wound up on digg, one of your readers would have to
refresh his page 50,000 times to break your access key.

So its perfectly ok if 10,000,000,000 readers or whatever refresh the
page 49,999 times EACH. The key will be fine.

Or at least thats the way I understand it.

Caching on the server is an option, but its really unnecessary. Google
is being really nice and offering to save you bandwidth. They're
pretty much offering unlimited free chart creation and hosting, as far
as i can see. Thats why i <3 Google :D They rock.

Sudhir

[aris]

so... is "user" the website or the request ip?

[barryhunter]

aris (and Sudhir),

see Uwe's reply already in this thread!
http://groups.google.com/group/google-chart-api/msg/b53c39a571fda6c8