Post-cutover cleanups to the back-end (#286)

* add the django-health-check module and a healthz endpoint

Signed-off-by: Irving Popovetsky <irving@popovetsky.com>

* Add a custom_sampler to ignore all traces to the health check

Signed-off-by: Irving Popovetsky <irving@popovetsky.com>

* make sure to add honeycomb auto-instrumentation for dev environments or else youll have a bad day trying to troubleshoot

Signed-off-by: Irving Popovetsky <irving@popovetsky.com>

* Correct the ALLOWED_HOST values for prod and staging

Signed-off-by: Irving Popovetsky <irving@popovetsky.com>

* linter did its thing

Signed-off-by: Irving Popovetsky <irving@popovetsky.com>

Author: irvingpop

poetry.lock

@@ -38,6 +38,7 @@ sentry-sdk = "^0.10.1"
 six = "^1.12"
 honeycomb-beeline = "^2.11.4"
 django-allow-cidr = "^0.3.1"
+django-health-check = "^3.12.1"
 
 [tool.poetry.dev-dependencies]
 bandit = "^1.6"

pyproject.toml

@@ -218,6 +218,21 @@ def worker_abort(worker):
     worker.log.info("worker received SIGABRT signal")
 
 
+def sampler(fields):
+    request_path = fields.get("request.path")
+    response_code = fields.get("response.status_code")
+
+    # never sample errors
+    if response_code and response_code >= 500:
+        return True, 1
+    else:
+        # never capture healthy health checks
+        if request_path == "/healthz":
+            return False, 0
+        # catchall
+        return True, 1
+
+
 # Added for Honeycomb instrumentation
 def post_worker_init(worker):
     worker.log.info("beeline initialization in process pid %s", worker.pid)
@@ -229,5 +244,7 @@ def post_worker_init(worker):
         beeline.init(
             writekey=os.getenv("HONEYCOMB_WRITEKEY"),
             dataset=os.getenv("HONEYCOMB_DATASET"),
+            service_name="backend",
+            sampler_hook=sampler,
             debug=False,
         )

src/gunicorn_config.py

@@ -35,6 +35,7 @@
         name="schema-swagger-ui",
     ),
     path("redoc/", schema_view.with_ui("redoc", cache_timeout=0), name="schema-redoc"),
+    path("healthz", include("health_check.urls")),
 ]
 
 ##############################################

src/operationcode_backend/urls.py

@@ -57,6 +57,10 @@
     # temp frontend apps
     "widget_tweaks",
     "snowpenguin.django.recaptcha2",
+    # django-health-check
+    # https://django-health-check.readthedocs.io/en/latest/
+    "health_check",  # required
+    "health_check.db",  # stock Django health checkers
 ]
 
 ROOT_URLCONF = "operationcode_backend.urls"

src/settings/components/base.py

@@ -13,3 +13,7 @@
     INSTALLED_APPS += ("debug_toolbar",)
 if "debug_toolbar.middleware.DebugToolbarMiddleware" not in MIDDLEWARE:
     MIDDLEWARE += ("debug_toolbar.middleware.DebugToolbarMiddleware",)
+
+# Honeycomb beeline auto-instrumentation
+if "beeline.middleware.django.HoneyMiddleware" not in MIDDLEWARE:  # noqa: F821
+    MIDDLEWARE += ("beeline.middleware.django.HoneyMiddleware",)  # noqa: F821

src/settings/environments/development.py

@@ -3,7 +3,7 @@
 from settings.components import config
 from settings.components.base import DATABASES
 
-ALLOWED_HOSTS = ["operationcode.org", "pybot.operationcode.org"]
+ALLOWED_HOSTS = ["api.operationcode.org"]
 DEBUG = False
 
 if config("EXTRA_HOSTS", default=""):

src/settings/environments/production.py

@@ -3,7 +3,7 @@
 from settings.components import config
 from settings.components.base import DATABASES
 
-ALLOWED_HOSTS = ["operationcode.org", "api.staging.operationcode.org"]
+ALLOWED_HOSTS = ["api.staging.operationcode.org"]
 DEBUG = False
 
 if config("EXTRA_HOSTS", default=""):