Merge branch 'master' into holmes-structured-output-support

Author: itisallgood

docs/configuration/holmesgpt/index.rst

@@ -18,7 +18,7 @@ Robusta can integrate with `Holmes GPT <https://github.com/robusta-dev/holmesgpt
 
 When available, AI based investigations can be launched in one of two ways:
 
-1. Click the ``Ask Holmes`` button in Slack. The AI investigation will be sent back as a new message.
+1. Click the ``Ask HolmesGPT`` button in Slack. The AI investigation will be sent back as a new message.
 
 .. image:: /images/robusta-holmes-investigation.png
     :width: 600px
@@ -223,6 +223,46 @@ Choose an AI provider below and follow the instructions:
 
         Run a :ref:`Helm Upgrade <Simple Upgrade>` to apply the configuration.
 
+    .. tab-item:: Multiple providers
+        :name: multiple-providers
+
+        Starting from version *0.22.1*, Robusta supports an alternative way to configure AI models: using a YAML dictionary in your Helm values file.
+
+        This method allows you to configure multiple models at once, each with its own parameters.
+
+        Update your Helm values (``generated_values.yaml`` file) with the following configuration.
+
+        When multiple models are defined, the Robusta UI will allow users to choose a specific model when initiating an AI-based investigation.
+
+        .. admonition:: Model info
+            :class: warning
+
+            When using multiple providers, the keys differ slightly from the single-provider case.
+
+        .. code-block:: yaml
+
+          enableHolmesGPT: true
+
+          holmes:
+            modelList: # sample configuration.
+              openai:
+                model: openai/gpt-4o
+                api_key: "{{ env.API_KEY }}"
+              azure-low-budget: 
+                model : azure/team-low-budget
+                api_base : <your-api-base> # fill in the base endpoint url of your azure deployment - e.g. https://my-org.openai.azure.com/
+                api_version : "2024-06-01"
+                api_key : "{{ env.AZURE_API_KEY }}" # you can load the values from an environment variable as well.
+                temperature: 0
+              bedrock-devops: 
+                model: bedrock/anthropic.claude-3-5-sonnet-20240620-v1:0 # your bedrock model.
+                aws_region_name: us-east-1
+                aws_access_key_id: "{{ env.AWS_ACCESS_KEY_ID }}" # you can load the values from an environment variable as well.
+                aws_secret_access_key: <your-aws-secret-access-key>
+                thinking: {"type": "enabled", "budget_tokens": 1024}
+
+        Run a :ref:`Helm Upgrade <Simple Upgrade>` to apply the configuration.
+
 Configuring HolmesGPT Access to SaaS Data
 ----------------------------------------------------
 

docs/configuration/holmesgpt/toolsets/prometheus.rst

@@ -24,7 +24,7 @@ Configuration
                 prometheus/metrics:
                     enabled: true
                     config:
-                        prometheus_url: http://<prometheus host>:9090
+                        prometheus_url: http://<prometheus host>:9090 # e.g. http://robusta-kube-prometheus-st-prometheus.default.svc.cluster.local:9090
                         headers:
                             Authorization: "Basic <base_64_encoded_string>"
 
@@ -41,7 +41,7 @@ Configuration
             prometheus/metrics:
                 enabled: true
                 config:
-                    prometheus_url: http://<prometheus host>:9090
+                    prometheus_url: http://<prometheus host>:9090 # e.g. http://robusta-kube-prometheus-st-prometheus.default.svc.cluster.local:9090
                     headers:
                         Authorization: "Basic <base_64_encoded_string>"
 
@@ -78,13 +78,11 @@ Below is the full list of options for this toolset:
 
 The best way to find the prometheus URL is to use "ask holmes". This only works if your cluster is live and already connected to Robusta.
 
-If not, follow these steps:
+If not, you can often find the prometheus URL by running the following command (several results may be shown - pick the best match):
 
-1. Run ``kubectl get services -n <monitoring-namespace>`` to list all services. Replace ``<monitoring-namespace>`` with the namespace where Prometheus is deployed. This is often ``monitoring`` or ``prometheus``. You can also run ``kubectl get services -A`` which will list all services in all namespaces.
-2. Identify which are the namespace and name of your Prometheus service. You can set up port forwarding to test if the service is correct and if Prometheus is reachable.
-3. Run ``kubectl describe service <service-name> -n <namespace>`` to get details about the service, including the cluster IP and port.
-4. Set the DNS or the cluster IP as well as the port to the configuration field ``prometheus_url`` as mentioned above.
+.. code-block:: bash
 
+    kubectl get svc --all-namespaces -o jsonpath='{range .items[*]}{.metadata.name}{"."}{.metadata.namespace}{".svc.cluster.local:"}{.spec.ports[0].port}{"\n"}{end}' | grep prometheus | grep -Ev 'operat|alertmanager|node|coredns|kubelet|kube-scheduler|etcd|controller' | awk '{print "http://"$1}'
 
 Capabilities
 ------------

docs/configuration/holmesgpt/toolsets/robusta.rst

@@ -3,7 +3,7 @@ Robusta :checkmark:`_`
 .. include:: ./_toolset_enabled_by_default.inc.rst
 
 By enabling this toolset, HolmesGPT will be able to fetch alerts metadata. It allows HolmesGPT to fetch information
-about specific issues when chatting using "Ask Holmes". This toolset is not necessary for Root Cause Analysis.
+about specific issues when chatting using "Ask HolmesGPT". This toolset is not necessary for Root Cause Analysis.
 
 Configuration
 -------------

src/robusta/core/model/base_params.py

@@ -83,7 +83,7 @@ class ResourceInfo(BaseModel):
 class HolmesParams(ActionParams):
 
     holmes_url: Optional[str]
-
+    model: Optional[str]
     @validator("holmes_url", allow_reuse=True)
     def validate_protocol(cls, v):
         if v and not v.startswith("http"):  # if the user configured url without http(s)
@@ -251,6 +251,7 @@ class HolmesWorkloadHealthChatParams(HolmesParams):
     conversation_history: Optional[list[dict]] = None
 
 
+
 class NamespacedResourcesParams(ActionParams):
     """
     :var name: Resource name

src/robusta/core/playbooks/internal/ai_integration.py

@@ -60,7 +60,8 @@ def ask_holmes(event: ExecutionBaseEvent, params: AIInvestigateParams):
             context=params.context if params.context else {},
             include_tool_calls=True,
             include_tool_call_results=True,
-            sections=params.sections
+            sections=params.sections,
+            model=params.model
         )
 
         if params.stream:
@@ -287,6 +288,7 @@ def holmes_issue_chat(event: ExecutionBaseEvent, params: HolmesIssueChatParams):
             conversation_history=params.conversation_history,
             investigation_result=params.context.investigation_result,
             issue_type=params.context.issue_type,
+            model=params.model
         )
         result = requests.post(f"{holmes_url}/api/issue_chat", data=holmes_req.json())
         result.raise_for_status()
@@ -336,7 +338,7 @@ def holmes_chat(event: ExecutionBaseEvent, params: HolmesChatParams):
     cluster_name = event.get_context().cluster_name
 
     try:
-        holmes_req = HolmesChatRequest(ask=params.ask, conversation_history=params.conversation_history)
+        holmes_req = HolmesChatRequest(ask=params.ask, conversation_history=params.conversation_history, model=params.model)
         result = requests.post(f"{holmes_url}/api/chat", data=holmes_req.json())
         result.raise_for_status()
         holmes_result = HolmesChatResult(**json.loads(result.text))
@@ -380,11 +382,12 @@ def holmes_workload_chat(event: ExecutionBaseEvent, params: HolmesWorkloadHealth
             ask=params.ask,
             conversation_history=params.conversation_history,
             workload_health_result=params.workload_health_result,
-            resource=params.resource
+            resource=params.resource,
+            model=params.model
         )
         result = requests.post(f"{holmes_url}/api/workload_health_chat", data=holmes_req.json())
         result.raise_for_status()
-        
+
         holmes_result = HolmesChatResult(**json.loads(result.text))
 
         finding = Finding(

src/robusta/core/reporting/holmes.py

@@ -20,6 +20,7 @@ class HolmesRequest(BaseModel):
     include_tool_calls: bool = False
     include_tool_call_results: bool = False
     sections: Optional[Dict[str, str]] = None
+    model: Optional[str] = None
 
 
 class HolmesConversationRequest(BaseModel):
@@ -35,6 +36,7 @@ class HolmesConversationRequest(BaseModel):
 class HolmesChatRequest(BaseModel):
     ask: str
     conversation_history: Optional[List[dict]] = None
+    model: Optional[str] = None
 
 
 class HolmesIssueChatRequest(HolmesChatRequest):

src/robusta/integrations/slack/sender.py

@@ -353,7 +353,7 @@ def __create_holmes_callback(self, finding: Finding) -> CallbackBlock:
 
         return CallbackBlock(
             {
-                "Ask Holmes": CallbackChoice(
+                "Ask HolmesGPT": CallbackChoice(
                     action=ask_holmes,
                     action_params=AIInvestigateParams(
                         resource=resource, investigation_type="issue", ask="Why is this alert firing?", context=context