diff --git a/deploy/k8s/base/chatbot/deployment.yaml b/deploy/k8s/base/chatbot/deployment.yaml
index 8a50a1ee..f7304677 100644
--- a/deploy/k8s/base/chatbot/deployment.yaml
+++ b/deploy/k8s/base/chatbot/deployment.yaml
@@ -11,6 +11,7 @@ spec:
     metadata:
       labels:
         app: crapi-chatbot
+    spec:
       containers:
       - name: crapi-chatbot
         image: crapi/crapi-chatbot:latest
diff --git a/deploy/k8s/base/deploy.sh b/deploy/k8s/base/deploy.sh
index a8e271fe..4cb988a6 100755
--- a/deploy/k8s/base/deploy.sh
+++ b/deploy/k8s/base/deploy.sh
@@ -12,3 +12,4 @@ kubectl apply -n crapi -f ./identity
 kubectl apply -n crapi -f ./community
 kubectl apply -n crapi -f ./workshop
 kubectl apply -n crapi -f ./web
+kubectl apply -n crapi -f ./chatbot
diff --git a/services/chatbot/src/chatbot_api.py b/services/chatbot/src/chatbot_api.py
index c199e8f7..bc3081d3 100644
--- a/services/chatbot/src/chatbot_api.py
+++ b/services/chatbot/src/chatbot_api.py
@@ -37,11 +37,10 @@ def document_loader():
             loader_cls=UnstructuredMarkdownLoader,
         )
         documents = loader.load()
-        app.logger.debug("Loaded %s documents", len(documents))
+        app.logger.debug("Loaded %s documents in db", len(documents))
         text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
         texts = text_splitter.split_documents(documents)
         embeddings = get_embeddings()
-        os.system("rm -rf ./db")
         db = Chroma.from_documents(texts, embeddings, persist_directory="./db")
         db.persist()
         retriever = db.as_retriever(search_kwargs={"k": target_source_chunks})