PINF-1150 stagger scheduler probes (#762)

Author: danielhoherd

Chart.lock

@@ -1,6 +1,6 @@
 dependencies:
 - name: airflow
-  repository: https://github.com/astronomer/apc-airflow/releases/download/oss-helm-chart/1.17.15-astro
-  version: 1.17.15-astro
-digest: sha256:4abf65ee0264a6868ce05ca0bb2fffda73889c7a5847cc6e9049fbe13edf3bd3
-generated: "2026-07-21T18:37:22.732846-04:00"
+  repository: https://github.com/astronomer/apc-airflow/releases/download/oss-helm-chart/1.17.16-astro
+  version: 1.17.16-astro
+digest: sha256:8fc6aa597647cda9e6238cab6983f521595fe9397966e329346d8f1c97178996
+generated: "2026-08-06T12:36:53.398345-04:00"

Chart.yaml

@@ -1,13 +1,13 @@
 # apiVersion v2 is Helm 3
 apiVersion: v2
 name: airflow
-version: 1.18.7
+version: 1.18.8
 description: Helm chart to deploy the Astronomer Platform Airflow module
 icon: https://airflow.apache.org/docs/apache-airflow/stable/_images/pin_large.png
 keywords:
   - astronomer
   - airflow
 dependencies:
   - name: airflow
-    version: 1.17.15-astro
-    repository: https://github.com/astronomer/apc-airflow/releases/download/oss-helm-chart/1.17.15-astro
+    version: 1.17.16-astro
+    repository: https://github.com/astronomer/apc-airflow/releases/download/oss-helm-chart/1.17.16-astro

bin/generate_circleci_config.py

@@ -8,7 +8,7 @@
 import yaml
 from jinja2 import Template
 
-git_root_dir = next(iter([x for x in Path(__file__).resolve().parents if (x / ".git").is_dir()]), None)
+git_root_dir = next(iter([x for x in Path(__file__).resolve().parents if (x / ".git").exists()]), None)
 metadata = yaml.safe_load((git_root_dir / "metadata.yaml").read_text())
 kube_versions = metadata["test_k8s_versions"]
 

tests/chart/test_airflow.py

@@ -171,3 +171,30 @@ def test_apiServer_startup_initialDelaySeconds_defaults(self, kube_version):
         assert len(docs) == 1
         c_by_name = get_containers_by_name(docs[0])
         assert c_by_name["api-server"]["startupProbe"]["initialDelaySeconds"] == 30
+
+    def test_scheduler_readinessprobe_and_livenessprobe_defaults_do_not_fire_in_lockstep(self, kube_version):
+        """PINF-1150: readinessProbe and livenessProbe run the identical exec
+        command (fork a process, import Airflow, round-trip the DB). If both fired
+        on the same initialDelaySeconds/periodSeconds, kubelet would trigger them
+        at the same instant every cycle, doubling the CPU/DB load from
+        health-checking alone at that moment -- readiness has no more timeout
+        headroom than liveness, so it would be the first to trip under contention.
+        This asserts values.yaml's override keeps the two offset, so a future edit
+        can't silently reintroduce the lockstep.
+        """
+        docs = render_chart(
+            kube_version=kube_version, show_only=["charts/airflow/templates/scheduler/scheduler-deployment.yaml"], values={}
+        )
+
+        assert len(docs) == 1
+        c_by_name = get_containers_by_name(docs[0])
+        liveness = c_by_name["scheduler"]["livenessProbe"]
+        readiness = c_by_name["scheduler"]["readinessProbe"]
+
+        assert liveness["timeoutSeconds"] == 30
+        assert liveness["initialDelaySeconds"] == 10
+        assert readiness["initialDelaySeconds"] == 40
+        assert liveness["periodSeconds"] == readiness["periodSeconds"] == 60
+        # Never coincide: the two schedules must not land on the same instant
+        # modulo their shared period.
+        assert (readiness["initialDelaySeconds"] - liveness["initialDelaySeconds"]) % liveness["periodSeconds"] != 0

values.yaml

@@ -56,6 +56,16 @@ airflow:
   scheduler:
     livenessProbe:
       timeoutSeconds: 30
+    # PINF-1150: readinessProbe and livenessProbe run the identical exec command
+    # (fork a process, import Airflow, round-trip the DB). apc-airflow's own chart
+    # defaults give both the same initialDelaySeconds/periodSeconds (10s/60s), so
+    # without this offset they fire at the exact same instant every 60s, doubling
+    # the CPU/DB load from health-checking alone at that moment -- readiness (the
+    # smaller timeout of the two) is the one that trips first under contention.
+    # 40s puts readiness exactly opposite liveness within the 60s cycle (10s and
+    # 40s, 30s apart), so the two health checks never coincide.
+    readinessProbe:
+      initialDelaySeconds: 40
     strategy:
       type: Recreate
     affinity: