can now pass in a query string `?statuses=x,y,z` to filter jobs based on
`Job.job_status`. Not passing in a status or passing in an empty string is
equivalent to selecting every filter type at once.
We want to show developers a log of notifications they’ve sent using the
API in the admin app. In order to indentify a notification it’s probably
helpful to know:
- who the notification was sent to (which we expose)
- when the notification was created (which we expose)
- which key was used to create the notification (which we expose, but
only as the `id` of the key)
Developers don’t see the `id` of the API key in the app anywhere, so
this isn’t useful information. Much more useful is the `type` and `name`
of the key. So this commit changes the schema to also return these.
This commit does some slightly hacky stuff with conftest because it
breaks a lot of other tests if the sample notification has a real API
key or an API key with a non-unique name.
Developers need visibility of what their integration is doing within
the app. This includes notifications sent with a test key.
This commit adds an optional, defaults-to-false parameter to include
notifications sent from a test API key when getting notifications.
- this allows us to send a notification to a provider by means of an API call
- This is in addition to the celery code.
- idea is that we can use this method to help speed up throughput by generating API traffic by node/lambda etc to supplement the celery code in times of high load.
This commit adds the `include_jobs` filter to the
`GET /services/…/notifications` endpoint. It defaults to `True` (ie show
all notifications) but makes it possible to only return notifications
created by _any_ API key.
This is so that we can show a log of all notifications sent through the
API in the admin app.
It does not expose this list to the public `GET /notifications` endpoint
because this would violate our rules about keys only being able to get
notifications created with keys of the same type.
- Thows a NoResultFound sqlalchemy exception
- Which causes a retry. This means we give it a few goes (5, max 5 hours) for the notification to appear.
- Should never happen, only if we get some task overlaps that are unusual that leads to tasks executed in an overlapping nature.
- Driven by the fact we won't know the type in the API call
- hence we need to load notification earlier , so pass it not the id through to the send task to avoid loading it twice.
accepts a page parameter to control what page of data
returns additional pagination fields in the response dict
* page_size: will always be 50. defined by Config.PAGE_SIZE
* total: the total amount of unpaginated records
* links: dict containing optionally prev, next, and last, links to
other relevant pagination pages
also cleaned up some test imports
- Aim to move the code that contacts providers into it's own module.
- Celery tasks now call this module to send to provider
- No exceptions caught in the new module. Celery tasks now use any exception to trigger a retry.
- tests moved about - new test directory for the new class, all tests from celery test module moved, excepting the retry logic.
- get all notifications by service
- template usage
- most recently used templates
Ensures that the dashboard shows no test key data. Supplements: https://github.com/alphagov/notifications-api/pull/677 which excludes CSV data. This branches from that so is dependant.
If you sign a token with a service ID that doesn’t exist (say, for
example, that you get service ID and API key mixed up) then you get
an error saying that “no API keys exist for the service”. This is wrong
because the service doesn’t even exist.
This commit adds:
- code to check if the service does exist
- a specific error message for this case
The check does mean an extra database call to look up the service.
However this only happens _after_ looping through all the API keys. So
it shouldn’t have a performance implication for anyone using a valid API
key.
If you create a token signed with a service ID that doesn’t exist, you
will get an error (as you should).
However we didn’t have a test that explicitly checks for this. This
commit adds one.
