When a platform admin user clicks on the button to add the
'upload_document' permission to a service, they should be taken to the
page to add a contact_link if the service does not already have one.
The contact link for the service now gets displayed on the service
settings page if the service has the 'upload_document' permission. There
are no changes to the appearance of the service settings page if a
service does not have the 'upload_document' permission.
Added a page which lets users with the 'manage_service' permission change the
contact link for their service. There are no links to this page yet
since only services using document download will need to set a contact
link.
At the moment branding is an undocumented feature. We get a bunch of
support tickets from teams asking its possible.
This commit:
- lets people know it’s possible, and what the options are
- is the first step towards making this process as self-service as
possible
In some cases we will be able to infer a user’s organisation from there
email address, and Google image search their logo. So the experience for
them is that they press a button and government just sorts it out for
you (also known as "the dream").
In other cases we will have to get back to people asking for a copy of
their logo, or to find out about their service, but this is what we have
to do at the moment anyway.
For both SMS senders and email reply to addresses this commit adds:
- a delete link
- a confirmation loop
It doesn’t let users delete:
- default SMS senders or reply to addresses (they always have to have
one)
- inbound numbers
It assumes that the API will allow updating of an attribute named
`active` on the respective database rows. It could work in a different
way. We can’t do complete deletion though because these will still be
keyed to notifications.
we reckon users will like to see gov reply-to email addresses because
it will improve their confidence in the email.
however, some services, for a few complex reasons, don't want a gov
reply to address. rather than add their specific domains to the
whitelist for signups etc, just make reply tos allowed from any domain.
We vet reply-tos before services go live anyway.
we were seeing isort produce different outputs locally and in docker -
this was due to it having different opinions about whether the tests
module (ie all our unit tests) is a first party (local) or third party
(pip installed) import. It's a first party import, so by defining this
in the setup.cfg isort settings, we can force it to be consistent
between environments.
Note: I don't know why it was different in the first place though
Most of the time spent by the admin app to generate a page is spent
waiting for the API. This is slow for three reasons:
1. Talking to the API means going out to the internet, then through
nginx, the Flask app, SQLAlchemy, down to the database, and then
serialising the result to JSON and making it into a HTTP response
2. Each call to the API is synchronous, therefore if a page needs 3 API
calls to render then the second API call won’t be made until the
first has finished, and the third won’t start until the second has
finished
3. Every request for a service page in the admin app makes a minimum
of two requests to the API (`GET /service/…` and `GET /user/…`)
Hitting the database will always be the slowest part of an app like
Notify. But this slowness is exacerbated by 2. and 3. Conversely every
speedup made to 1. is multiplied by 2. and 3.
So this pull request aims to make 1. a _lot_ faster by taking nginx,
Flask, SQLAlchemy and the database out of the equation. It replaces them
with Redis, which as an in-memory key/value store is a lot faster than
Postgres. There is still the overhead of going across the network to
talk to Redis, but the net improvement is vast.
This commit only caches the `GET /service` response, but is written in
such a way that we can easily expand to caching other responses down the
line.
The tradeoff here is that our code is more complex, and we risk
introducing edge cases where a cache becomes stale. The mitigations
against this are:
- invalidating all caches after 24h so a stale cache doesn’t remain
around indefinitely
- being careful when we add new stuff to the service response
---
Some indicative numbers, based on:
- `GET http://localhost:6012/services/<service_id>/template/<template_id>`
- with the admin app running locally
- talking to Redis running locally
- also talking to the API running locally, itself talking to a local
Postgres instance
- times measured with Chrome web inspector, average of 10 requests
╲ | No cache | Cache service | Cache service and user | Cache service, user and template
-- | -- | -- | -- | --
**Request time** | 136ms | 97ms | 73ms | 37ms
**Improvement** | 0% | 41% | 88% | 265%
---
Estimates of how much storage this requires:
- Services: 1,942 on production × 2kb = 4Mb
- Users: 4,534 on production × 2kb = 9Mb
- Templates: 7,079 on production × 4kb = 28Mb
Adding a ‘testing’ template it not enough. It needs to have some real
looking content, so that we can:
- work out what a service is doing
- assess whether that’s a reasonable (ie meeting the terms of use) thing
to be doing with Notify
At the moment we’re having to go back to services quite a lot when they
request to go live and ask them for this stuff.
`<h1>`s should be unique across the site. This page’s `<h1>` matches
that of the previous page (the one with the checklist).
This commit re-titles it to:
- be unique
- more accurately describe the content of the page
This question was designed to make people feel like it was OK to submit
their request without getting the MoU signed. We reckoned that this was
the fastest way of getting their service live (because the MoU is the
bit that’s most likely to slow them down).
We now have a better way of telling people:
- if they’ve signed the MoU already
- or to contact us if they haven’t (which is what the majority of teams
seem to do now)
We were never actually using the answer to this question – we were still
checking for every service whether they had it signed.
So this commit removes this now-redundant question.
We were counting users who had the `manage_settings` permission. This
is the old name for it, therefore there would never be any users with
this permission, so the tick would never go green.
The new name for the permission is `manage_service`. This commit fixes
the error, and adds an extra safeguard against something like this
happening again.
This makes it easier to write a good message in the request to go live
submission. And encapsulating it in the `GovernmentDomain` class keeps
the view nice and clean.
When we process a go live request it’s a pain to have to go fishing in
the Google Drive to see if there’s a signed agreement in there. This
should make it easier when we know there’s already an agreement.
The free allowance affect the number of free text messages a services get per yer.
This allowance is being set properly when an organisation is created by not updated.
This PR updates the free allowance when the organisation type is updated.
The free allowance can also be changed if service has an exceptional free allowance.
Having SMS senders that start with 00 can cause issues with Firetext due
to Firetext's validation rules, so we shouldn't allow SMS senders to start
with 00.
Firetext treats a double 00 at the start of the senderID as an international
prefix, so removes them. A sender of 00447876574016 would become 447876574016.
Under Firetext's validation rules, an SMS sender of five 0s (00000) would
become 4400. This is because the first 00 are removed (as the international
prefix). The third 0 is seen as the start of a phone number, and becomes 44,
leaving the final 00 = 4400.
Done using isort[1], with the following command:
```
isort -rc ./app ./tests
```
Adds linting to the `run_tests.sh` script to stop badly-sorted imports
getting re-introduced.
Chosen style is ‘Vertical Hanging Indent’ with trailing commas, because
I think it gives the cleanest diffs, eg:
```
from third_party import (
lib1,
lib2,
lib3,
lib4,
)
```
1. https://pypi.python.org/pypi/isort
We require that a user has a real reply-to email address before going
live. We can partially automate this by at least telling users who
haven’t done this.
This only applies for users that have email templates; we shouldn’t
bother users who aren’t going to send emails about this.
We need users to have created some templates before they go live, so we
can see what kind of messages they intend to send.
We can do this automatically based on the work done in
https://github.com/alphagov/notifications-admin/pull/1892
One of the things that we want to check before a service goes live is
that they have at least two team members with the manage service
permission. Anyone who can make a request to go live has this
permission, so that means one additional user is needed. This is what we
can automatically communicate to the user.
Under the hood this makes use of the logic added in
https://github.com/alphagov/notifications-admin/pull/1891
Users who have the ‘manage API keys’ permission can see the settings
page. But they don’t have permission to request to go live.
At the moment they can still see the link, though clicking it gives them
a 403 error. This commit changes it so that they can’t see the link, and
tells them who they should speak to about going live (their manager).
A lot of users aren’t reading or paying attention to the checklist on
the request to go live page. We think that we can get more people to
read it by putting it on its own page, where users won’t jump straight
to filling in the form.
This will, later on, let us make this page smarter by automatically
detecting if they’ve done the necessary things.
Notifications-api now needs the service_id to check the uniqueness of
the service name when trying to change it. This is to allow a user to
successfully change the pluralization or the case of their service name.
We have a team who want their (short) web address as the text message
sender. This commit updates the validation of text message senders to
allow `.` as a valid character, which is currently blocking them from
doing this.
We can be fairly confident this works because:
- the team are sending large volumes of messages already with their
existing provider
- we’ve tested it with all combinations of
- both our text message providers
- an Android phone and n iPhone
I personally think it’s more robust to have a test also cover the
counter-fact.
And it’s easy to understand what’s going on if you can see the valid and
invalid examples side by side.
Using a separate validator class to check for appropriate characters in
a text message sender means that we’re not doing this validation in a
different way from the other checks (length and required). So the code
is cleaner.
Users can choose whether to have these features on or off now. And
a platform admin can still use the same controls to switch the features
on or off on behalf of a user. There’s no need for the separate buttons.
Our support ticket analysis shows that the most common action request
after going live is turning on letters.
We just do this for any team that requests it – there’s no gatekeeping.
So we should just allow people to make the change themselves.
This will be a better experience for our users, and less work for us.
The design of the page replicates roughly what we have for international
text messaging.
One of the things that came out of the support analysis was that people
were asking how much inbound SMS costs. There wasn’t a significant
volume of these requests, but the fix seems low-effort and
non-disruptive enough that we should do it.
Content by Thom.
