The jsonify decorator parses request payload as JSON
and puts it in request.json. The payload would normally
be a complex object, but if a client sends, let's say,
a single integer, then request.json is a python int.

The authorize decorator looks for API key first in request
headers, then in request body. It expects the request
body to be a complex object.

This commit changes adds the following validation rule in
the jsonify decorator: if request body is not empty, it
*must* parse as JSON, and the root element of the parsed
document *must* be a dict.

Fixes: #574

Previous retry logic was:
- max 3 tries
- every try times out after 5 seconds

The new retry logic is:
- max 3 tries
- every try times out after 10 seconds
- if the first two tries have used > 10 seconds, don't
  do the third try

cc: #569

Fixes: #566

Fixes: #565

Also, split Dockerfile into two stages, so rust
and other build dependencies don't end up in the final image.

Note cryptography has binary wheels for various architectures,
but unfortunately not for 32-bit arm. And, starting from v35.0.0,
cryptography requires rust to build from source.

1. Drop API support for GET, DELETE requests with a request body.
Healthchecks had an undocumented quirk where you could authenticate a
GET or DELETE request by putting a '{"api_key":"..."}' in request body.
This commit removes this feature.

Note: POST requests can still authenticate either by sending
a X-Api-Key header, or by putting a "api_key" key in request body.
GET and DELETE requests can now only authenticate with the
request header.

2. Add missing @csrf_exempt annotations in API views
When client sends a HTTP POST request to a GET-only endpoint,
the server is supposed to respond with "405 Method Not Allowed".
Due to CSRF checking, a couple endpoints were responding with
"403 Forbidden" instead. Adding @csrf_exempt annotations fixes
the problem.

cc: #552

Fixes: #77

This fixes a security issue:
- attacker can crafts a redirect URL to an external site
- attacker gets victim to click on it
- victim logs in
- after login, Healthchecks redirects victim to the external site

The _allow_redirect function now additionally
requires the redirect URL is relative (has no scheme or domain).

Fixes: #548

Fixes: #354

Fixes: #323

Fixes: #539

PagerDuty Connect is deprecated and will be discontinued.
It is replaced by PagerDuty Simple Install Flow (see
README for setup instructions).

Fixes: #512

Primary sort key: projects with overall_status=down go first
Secondary sort key: project's name

Fixes: #497

Profile.next_nag_date tracks when the next hourly/daily reminder
should be sent. Normally, sendalerts sets this field when
a check goes down, and sendreports clears it out whenever
it is about to send a reminder but realizes all checks are up.

The problem: sendalerts can set next_nag_date to a non-null
value, but it does not clear it out when all checks are up.
This can result in a hourly/daily reminder being sent out
at the wrong time. Specific example, assuming hourly reminders:

13:00: Check A goes down. next_nag_date gets set to 14:00.
13:05: Check A goes up. next_nag_date remains set to 14:00.
13:55: Check B goes down. next_nag_date remains set to 14:00.
14:00: Healthchecks sends a hourly reminder, just 5 minutes
       after Check B going down. It should have sent the reminder
       at 13:55 + 1 hour = 14:55

The fix: sendalerts can now both set and clear the next_nag_date
field. The main changes are in Project.update_next_nag_dates()
and in Profile.update_next_nag_date(). With the fix:

13:00: Check A goes down. next_nag_date gets set to 14:00.
13:05: Check A goes up. next_nag_date gets set to null.
13:55: Check B goes down. next_nag_date gets set to 14:55.
14:55: Healthchecks sends a hourly reminder.

Fixes: #486