diff --git a/pyvatti/src/pyvatti/helm.py b/pyvatti/src/pyvatti/helm.py
index e0cff31..0998c0b 100644
--- a/pyvatti/src/pyvatti/helm.py
+++ b/pyvatti/src/pyvatti/helm.py
@@ -85,6 +85,7 @@ class GlobflowParams(BaseModel):
     outdir: str
     config_application: str
     config_crypt4gh: str
+    secret_key: str
 
 
 class Secrets(BaseModel):
diff --git a/pyvatti/src/pyvatti/jobstates.py b/pyvatti/src/pyvatti/jobstates.py
index 22be90d..03d1d3e 100644
--- a/pyvatti/src/pyvatti/jobstates.py
+++ b/pyvatti/src/pyvatti/jobstates.py
@@ -2,8 +2,8 @@
 
 
 class States(str, enum.Enum):
-    REQUESTED = "requested"
-    CREATED = "created"
-    DEPLOYED = "deployed"
-    FAILED = "failed"
-    SUCCEEDED = "succeeded"
+    REQUESTED = "Requested"
+    CREATED = "Created"
+    DEPLOYED = "Deployed"
+    FAILED = "Failed"
+    SUCCEEDED = "Succeeded"
diff --git a/pyvatti/src/pyvatti/notifymodels.py b/pyvatti/src/pyvatti/notifymodels.py
index 1d8a576..2cee576 100644
--- a/pyvatti/src/pyvatti/notifymodels.py
+++ b/pyvatti/src/pyvatti/notifymodels.py
@@ -54,7 +54,7 @@ class SeqeraLog(BaseModel):
     Parse the seqera status to a job machine state:
 
     >>> log.get_job_state()
-    <States.SUCCEEDED: 'succeeded'>
+    <States.SUCCEEDED: 'Succeeded'>
 
     A job can take time to start. During this time you'll get empty responses from the Seqera API:
 
@@ -103,23 +103,15 @@ def get_job_state(self) -> Optional[States]:
         return state
 
 
-class BackendEvents(str, enum.Enum):
-    """Events recognised by the backend"""
-
-    STARTED = "started"
-    ERROR = "error"
-    COMPLETED = "completed"
-
-
 class BackendStatusMessage(BaseModel):
     """A message updating the backend about job state
 
     >>> from datetime import datetime
-    >>> d = {"run_name": "INTP123456", "utc_time": datetime(1999, 12, 31), "event": BackendEvents.COMPLETED}
+    >>> d = {"run_name": "INTP123456", "utc_time": datetime(1999, 12, 31), "event": States.DEPLOYED }
     >>> BackendStatusMessage(**d).model_dump_json()
-    '{"run_name":"INTP123456","utc_time":"1999-12-31T00:00:00","event":"completed"}'
+    '{"run_name":"INTP123456","utc_time":"1999-12-31T00:00:00","event":"Deployed"}'
     """
 
     run_name: str
     utc_time: PastDatetime
-    event: BackendEvents
+    event: States
diff --git a/pyvatti/src/pyvatti/pgsjob.py b/pyvatti/src/pyvatti/pgsjob.py
index 383f7f2..4a98f6d 100644
--- a/pyvatti/src/pyvatti/pgsjob.py
+++ b/pyvatti/src/pyvatti/pgsjob.py
@@ -14,7 +14,7 @@
 from pyvatti.config import Settings, K8SNamespace
 from pyvatti.jobstates import States
 from pyvatti.messagemodels import JobRequest
-from pyvatti.notifymodels import SeqeraLog, BackendStatusMessage, BackendEvents
+from pyvatti.notifymodels import SeqeraLog, BackendStatusMessage
 
 from pyvatti.resources import GoogleResourceHandler, DummyResourceHandler
 
@@ -35,7 +35,7 @@ class PolygenicScoreJob(Machine):
     On instantiation the default job state is requested:
 
     >>> job.state
-    <States.REQUESTED: 'requested'>
+    <States.REQUESTED: 'Requested'>
 
     Normally creating a resource requires some parameters from a message, but not in dry run mdoe:
 
@@ -56,7 +56,7 @@ class PolygenicScoreJob(Machine):
 
     >>> job.trigger("succeed") # doctest: +ELLIPSIS
     Sending state notification: States.SUCCEEDED
-    msg='{"run_name":"INT123456","utc_time":...,"event":"completed"}' prepared to send to pipeline-notify topic (PYTEST RUNNING)
+    msg='{"run_name":"INT123456","utc_time":...,"event":"Succeeded"}' prepared to send to pipeline-notify topic (PYTEST RUNNING)
     Deleting all resources: INT123456
     ...
 
@@ -72,7 +72,7 @@ class PolygenicScoreJob(Machine):
     >>> bad_job = PolygenicScoreJob("INT789123", dry_run=True)
     >>> bad_job.trigger("error")  # doctest: +ELLIPSIS
     Sending state notification: States.FAILED
-    msg='{"run_name":"INT789123","utc_time":"...","event":"error"}' prepared to send to pipeline-notify topic (PYTEST RUNNING)
+    msg='{"run_name":"INT789123","utc_time":"...","event":"Failed"}' prepared to send to pipeline-notify topic (PYTEST RUNNING)
     Deleting all resources: INT789123
     ...
 
@@ -125,12 +125,6 @@ class PolygenicScoreJob(Machine):
         States.SUCCEEDED: "succeed",
         States.DEPLOYED: "deploy",
     }
-    # map from states to events recognised by the backend
-    state_event_map: ClassVar[dict] = {
-        States.FAILED: BackendEvents.ERROR,
-        States.SUCCEEDED: BackendEvents.COMPLETED,
-        States.DEPLOYED: BackendEvents.STARTED,
-    }
 
     def __init__(
         self, intp_id: str, settings: Optional[Settings] = None, dry_run: bool = False
@@ -171,6 +165,7 @@ def __init__(
 
     def handle_error(self, event):
         logger.warning(f"Exception raised for {self.intp_id}")
+        logger.warning(event.error)
         if isinstance(event.error, MachineError):
             logger.warning(f"Couldn't trigger error state for {self.intp_id}")
             raise event.error
@@ -197,9 +192,8 @@ def notify(self, _: EventData):
         """Notify the backend about the job state"""
         # bootstrap_server_host: str, bootstrap_server_port: int
         logger.info(f"Sending state notification: {self.state}")
-        event: str = self.state_event_map[self.state]
         msg: str = BackendStatusMessage(
-            run_name=self.intp_id, utc_time=datetime.now(), event=event
+            run_name=self.intp_id, utc_time=datetime.now(), event=self.state
         ).model_dump_json()
         if "pytest" in sys.modules:
             logger.info(
diff --git a/pyvatti/tests/test_render.py b/pyvatti/tests/test_render.py
index e1d7d7a..191e105 100644
--- a/pyvatti/tests/test_render.py
+++ b/pyvatti/tests/test_render.py
@@ -19,7 +19,7 @@ def test_render(message):
     """Test a values.yaml helm file is templated and rendered correctly"""
     settings = Settings(
         TOWER_TOKEN="test",
-        TOWER_WORKSPACE="test",
+        TOWER_WORKSPACE=000000,
         GLOBUS_DOMAIN="https://example.com",
         GLOBUS_CLIENT_ID="test",
         GLOBUS_CLIENT_SECRET="test",