Re-enable health check that includes crunchydb cluster check

api/app/health.py

@@ -7,14 +7,14 @@
 logger = logging.getLogger(__name__)
 
 
-def patroni_cluster_health_check():
+def crunchydb_cluster_health_check():
     """ Makes call to Patroni cluster namespace in Openshift to retrieve the statuses of all
     individual Patroni pods, then re-formats response """
     parts = [
         config.get('OPENSHIFT_BASE_URI'),
         config.get('OPENSHIFT_NAMESPACE_API'),
         config.get('PROJECT_NAMESPACE'),
-        'statefulsets/',
+        'postgresclusters/',
         config.get('PATRONI_CLUSTER_NAME')
     ]
     # form URL by concatenating all substrings in parts[], making sure there's exactly 1 / between each part
@@ -24,12 +24,12 @@ def patroni_cluster_health_check():
     }
     resp = requests.get(url, headers=header, timeout=10)
     resp_json = resp.json()
-    # NOTE: In Openshift parlance "replica" refers to how many of one pod we have, in Patroni, a "Replica"
-    # refers to a read only copy of of the Leader.
+    # NOTE: In Openshift parlance "replica" refers to how many of one pod we have, in CrunchyDB's managed
+    # Patroni, a "Replica" refers to a read only copy of of the Leader.
     # Get the number of pods that are ready:
-    ready_count = resp_json.get('status').get('readyReplicas')
+    ready_count = resp_json.get('status').get('instances')[0].get('readyReplicas')
     # Get the number of pods we expect:
-    replica_count = resp_json.get('status').get('replicas')
+    replica_count = resp_json.get('status').get('instances')[0].get('replicas')
     if ready_count > 1:
         # It's actually a bit more complicated than this.
         # There are a number of scenarios that are ok:

api/app/main.py

@@ -12,7 +12,7 @@
 from app.percentile import get_precalculated_percentiles
 from app.auth import authentication_required, audit
 from app import config
-# from app import health
+from app import health
 from app import hourlies
 from app.rocketchat_notifications import send_rocketchat_notification
 from app.routers import (fba, forecasts, weather_models, c_haines, stations, hfi_calc,
@@ -123,11 +123,10 @@ async def get_health():
     """ A simple endpoint for Openshift Healthchecks.
     It's assumed that if patroni is ok, then all is well.  """
     try:
-        # TODO reenable
-        # health_check = health.patroni_cluster_health_check()
+        health_check = health.crunchydb_cluster_health_check()
 
-        # logger.debug('/health - healthy: %s. %s',
-        #              health_check.get('healthy'), health_check.get('message'))
+        logger.debug('/health - healthy: %s. %s',
+                     health_check.get('healthy'), health_check.get('message'))
 
         # Instantiate the CFFDRS singleton. Binding to R can take quite some time...
         cffdrs_start = perf_counter()
@@ -138,7 +137,7 @@ async def get_health():
         if delta > 0.1:
             logger.info('%f seconds added by CFFDRS startup', delta)
 
-        return {"message": "API healthy", "healthy": True}
+        return health_check
     except Exception as exception:
         logger.error(exception, exc_info=True)
         raise