Configure Toolforge health-check-path

With this, `webservice restart` should perform a no-downtime rolling
restart, where Kubernetes only terminates the old pod once the new one
is ready. The tool might also be restarted automatically if the health
check starts failing (though I’m not sure how often this would happen
with the “blank” health check we configure here, which really only
confirms uWSGI is running).

Add a new /healthz route, and put it in the service.template; also,
since this route is going to be hit by a lot of requests, disable
logging for it. And since this means we now have an uwsgi.ini file
anyway, we might as well also configure automatic log rotation, as
documented at [1] – though I prefer a slightly larger maximum file size
than suggested there.

[1]: https://wikitech.wikimedia.org/wiki/Help:Toolforge/Web/Python#Logs

Bug: T341919

{{cookiecutter.tool_identifier}}/app.py

@@ -219,6 +219,11 @@ def logout(){% if cookiecutter.set_up_mypy == "True" %} -> RRV{% endif %}:
     return flask.redirect(flask.url_for('index'))
 
 
+@app.route('/healthz')
+def health(){% if cookiecutter.set_up_mypy == "True" %} -> RRV{% endif %}:
+    return ''
+
+
 def full_url(endpoint{% if cookiecutter.set_up_mypy == "True" %}: str{% endif %}, **kwargs){% if cookiecutter.set_up_mypy == "True" %} -> str{% endif %}:
     scheme = flask.request.headers.get('X-Forwarded-Proto', 'http')
     return flask.url_for(endpoint, _external=True, _scheme=scheme, **kwargs)

{{cookiecutter.tool_identifier}}/service.template

@@ -2,3 +2,4 @@
 # Provide default arguments for `webservice` commands for this tool.
 backend: kubernetes
 type: python3.11
+health-check-path: /healthz

{{cookiecutter.tool_identifier}}/uwsgi.ini

@@ -0,0 +1,6 @@
+[uwsgi]
+# rotate ~/uwsgi.log after 100 MiB
+log-maxsize = 104857600
+# don’t log the /healthz health-check endpoint to ~/uwsgi.log
+# (it’s included in Kubernetes startup and liveness probes so it gets lots of requests)
+route = ^/healthz donotlog: