Takes the number of emails and SMS fragments sent from:
https://github.com/alphagov/notifications-api/pull/273
Using these numbers it’s possible to show:
- how much of your allowance is left
- or how much you have spent
For now the allowance and rates are hard coded.
Only for users that have manage service.
If they API returns no statistics for a given time period we should
assume that this is equivalent to 0. This means that the template can
always rely on the dictionary having the same keys.
Requested, delivered and failed are the three states stored in the
notification statistics table. They are not discrete, eg a message can
be counted in delivered and failed.
`requested` is incremented *when a notification is created*, and has no
chance of being incremented twice for the same notification.
The template statistics are incremented *when a notification is created
only*.
Therefore the only way to make the numbers line up is to count:
- messages sent as being `requested`
- failure rate being `failed`/`requested` *not*
`failed`/`failed`+`delivered`
The dashboard should only talk about notifications that are *successful* or
*failed*. It should not count notifications that are still queued.
This will stop:
- a situation like ‘0 emails, 14.1% failed’
- the big numbers and the template statistics numbers not adding up to
the same total
> We show the last weeks template usage on the dashboard, which is
> great, but if you're looking for longer term trends, you're out of
> luck...
> So, let's let you see more on a more detailed page (linked from the
> dashboard). Initially this should just show you all templates that you
> have used ever and the count for each. Order same as dashboard, most
> popular first.
https://www.pivotaltracker.com/story/show/117614585
Previous the table of templates on the dashboard was for all time.
This commit uses the `limit_days` parameter of the API endpoint to only
show template usage from the last 7 days, aligning with the big numbers
shown above.
We want to align all our stats to be for the last 7 days.
This means summing up the stats response from the API to make the Big
Number. Previously the big number was counting sent notifications as
successful. This commit changes it to only look at delivered
notifications.
Right now, the API doesn’t have a way of filtering to only show the last
7 days. So for the moment the dashboard will show statistics for all
time.
The upshot of this is that we can link from the dashboard to the
activity page when there are failures.
When you add a new service, it’s probably the one you want to do stuff
with.
When you get invited, the service you’ve been invited to is probably the
one you want to use.
This commit adds the ID of the new service or service you’ve been
invited to to the session.
There’s no content in the tour that’s specific to a service. And since
we can now take a pretty good guess at what service you last used, or
which service we should send you to if you only have one service,
there’s no need to make the URLs for the tour service-specific.
This also means that you don’t need to be logged in to see the tour
pages, and we have no good reason to only restrict these pages to users
with accounts.
https://www.pivotaltracker.com/story/show/116960703
Template stats should show the most-used template first.
This commit:
- re-writes the `aggregate_usage` function to use `itertools.groupby`,
which can do aggregation, and can return data in a structure that’s
easy to sort on
- uses generators so that we’re not keeping lots of rows of template
stats in memory
https://www.pivotaltracker.com/story/show/117348893
A new permission has been added, view_activity, to resolve this issue.
Another pull request in notifications-admin will be required to update all users with a default permission of view_activity.
https://www.pivotaltracker.com/story/show/116052359
- add absolute numbers for failures
- make percentages accurate to 1 decimal place (50.0%)
- make error colour appear if failures go above 3.0%
- make error colour boolean (don’t interpolate between the colours)
This commit adds two new sections to the dashboard
1. A banner telling you about trial mode, including a count of how many messages
you have left today, which is a restriction of trial mode
2. Panels with counts of how many emails and text messages have been sent in a
day, plus the failure rates for each
It does **not**:
- link through to any further information about what trial mode is (coming
later)
- link through to pages for the failure rates (coming later)
- change the ‘recent jobs’ section to ‘recent notifications’
Based on discussion with Pete.
Make the blue banner an ‘important’ banner (copied from Register to Vote, used
because it’s not as boxy and fits on the page better).
Remove the back button because you haven’t changed any data yet. If you need to
go back you can just press back or start again.
Make the filename stand out more.
Remove the ‘download example’ link. Will need to revist the best way of doing
this.
Make text messages consistently 2/3rd width.
