Merge branch 'master' into ROB-2060-slack-sink-proxy

RoiGlinik · RoiGlinik · commit ba0b4ff3c56d · 2025-09-21T13:57:20.000+03:00
diff --git a/docs/configuration/holmesgpt/getting-started.rst b/docs/configuration/holmesgpt/getting-started.rst
@@ -43,6 +43,10 @@ Use Robusta's hosted AI service with GPT-4o:
 
 That's it! HolmesGPT will now analyze your alerts automatically.
 
+.. note::
+
+   When exploring HolmesGPT documentation, focus on **Robusta Helm chart configuration** sections rather than CLI installation. Robusta users should follow the Robusta Helm Chart based configuration examples for data sources and advanced settings.
+
 Test Your Setup
 ---------------
 
@@ -167,13 +171,13 @@ Common Issues
    - Consider using Robusta AI for unlimited investigations
 
 **Analysis seems incomplete?**
-   - Enable additional data sources in `HolmesGPT configuration <https://holmesgpt.dev/data-sources/builtin-toolsets/>`_
+   - Enable additional data sources in `HolmesGPT data sources <https://holmesgpt.dev/data-sources/builtin-toolsets/>`_ (follow Helm chart configuration examples)
    - Ensure Prometheus is configured for metrics analysis
    - Check that pod logs are accessible
 
 Next Steps
 ----------
 
 * :doc:`main-features` - See what HolmesGPT can do
-* `Configure Data Sources <https://holmesgpt.dev/data-sources/builtin-toolsets/>`_ - Add more context for better analysis
-* `Advanced Configuration <https://holmesgpt.dev>`_ - Customize HolmesGPT behavior
+* `Configure Data Sources <https://holmesgpt.dev/data-sources/builtin-toolsets/>`_ - Add more context for better analysis (use Helm chart configuration)
+* `Helm Configuration Reference <https://holmesgpt.dev/reference/helm-configuration/>`_ - Advanced HolmesGPT Helm settings
diff --git a/docs/configuration/holmesgpt/holmesgpt-docs.rst b/docs/configuration/holmesgpt/holmesgpt-docs.rst
@@ -0,0 +1,27 @@
+HolmesGPT Documentation
+========================
+
+For comprehensive HolmesGPT documentation, please visit the official HolmesGPT documentation site at `holmesgpt.dev <https://holmesgpt.dev>`_.
+
+Configuring HolmesGPT with Robusta
+-----------------------------------
+
+When configuring HolmesGPT with Robusta, follow the **Robusta Helm Chart configuration method** described in the HolmesGPT documentation rather than standalone CLI installation.
+
+Key points:
+
+* Use the ``enableHolmesGPT: true`` setting in your Robusta Helm values
+* Configure data sources and advanced settings using the Helm chart configuration examples provided in the HolmesGPT docs
+* Refer to the `Helm Configuration Reference <https://holmesgpt.dev/reference/helm-configuration/>`_ for advanced HolmesGPT settings specific to Robusta deployments
+
+Quick Links
+-----------
+
+* `HolmesGPT Main Documentation <https://holmesgpt.dev>`_
+* `Data Sources Configuration <https://holmesgpt.dev/data-sources/builtin-toolsets/>`_
+* `Helm Chart Configuration Reference <https://holmesgpt.dev/reference/helm-configuration/>`_
+
+For Robusta-specific HolmesGPT setup instructions, see:
+
+* :doc:`getting-started` - Quick setup guide for HolmesGPT with Robusta
+* :doc:`main-features` - Overview of AI analysis features
diff --git a/docs/configuration/holmesgpt/main-features.rst b/docs/configuration/holmesgpt/main-features.rst
@@ -74,5 +74,5 @@ Next Steps
 ----------
 
 * :doc:`getting-started` - Set up HolmesGPT in 5 minutes
-* `HolmesGPT Documentation <https://holmesgpt.dev>`_ - Advanced configuration and customization
-* `Available Data Sources <https://holmesgpt.dev/data-sources/builtin-toolsets/>`_ - See all supported integrations
+* `Available Data Sources <https://holmesgpt.dev/data-sources/builtin-toolsets/>`_ - See all supported integrations (use Robusta Helm Chart configuration method)
+* `Helm Configuration Reference <https://holmesgpt.dev/reference/helm-configuration/>`_ - Advanced HolmesGPT settings for Robusta deployments
diff --git a/docs/index.rst b/docs/index.rst
@@ -74,7 +74,7 @@
 
    configuration/holmesgpt/main-features
    configuration/holmesgpt/getting-started
-   Configuring HolmesGPT <https://holmesgpt.dev>
+   HolmesGPT Docs <configuration/holmesgpt/holmesgpt-docs>
 
 .. toctree::
    :maxdepth: 4
diff --git a/docs/notification-routing/configuring-sinks.rst b/docs/notification-routing/configuring-sinks.rst
@@ -90,6 +90,22 @@ Description of each option:
 | *parameters*     | Slack and webhook_url for MSTeams                       |                                                          |                                               |
 +------------------+---------------------------------------------------------+----------------------------------------------------------+-----------------------------------------------+
 
+Ignoring Sinks Initialization Errors
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+By default, when Robusta fails to initialize any of the Sinks, it will not start.
+
+On some scenarios, you may want to ignore Sinks initialization errors.
+
+For example, if Robusta is not allowed to connect to Slack, but you still want to receive notifications on the Robusta UI.
+
+In order to enable that, add the below to ``globalConfig`` in your ``generated_values.yaml`` file:
+
+.. code-block:: yaml
+
+    globalConfig:
+      continue_on_sink_errors: True
+
 Learn More
 ^^^^^^^^^^^^
 
diff --git a/docs/setup-robusta/installation/all-in-one-installation.rst b/docs/setup-robusta/installation/all-in-one-installation.rst
@@ -26,4 +26,5 @@ Prerequisites
 Next Steps
 ---------------------------------
 
-:ref:`See how Robusta improves Prometheus <builtin-alert-enrichment>`.
+* :ref:`See how Robusta improves Prometheus <builtin-alert-enrichment>`
+* :ref:`Learn about AI-powered alert analysis <AI Analysis>`
diff --git a/docs/setup-robusta/installation/extend-prometheus-installation.rst b/docs/setup-robusta/installation/extend-prometheus-installation.rst
@@ -29,5 +29,6 @@ Next Steps
 
 Integrate Robusta with AlertManager:
 
-* :ref:`Follow a guide to integrate AlertManager <Integrating with Prometheus>`.
-* :ref:`See the features you'll gain by integrating AlertManager <Enhanced Prometheus Alerts>`.
+* :ref:`Follow a guide to integrate AlertManager <Integrating with Prometheus>`
+* :ref:`See the features you'll gain by integrating AlertManager <Enhanced Prometheus Alerts>`
+* :ref:`Configure AI analysis with your Prometheus data <AI Analysis>`
diff --git a/docs/setup-robusta/installation/standalone-installation.rst b/docs/setup-robusta/installation/standalone-installation.rst
@@ -7,7 +7,7 @@ Use HolmesGPT without Prometheus
 
 *Estimated time: 5 minutes*
 
-Robusta's AI Agent works with many monitoring tools beyond Prometheus - including Datadog, New Relic, PagerDuty, and more. This installation method is ideal when you already have monitoring infrastructure in place and want to enhance it with Robusta's AI-powered investigation and automation capabilities.
+Robusta's AI Agent works with many monitoring tools beyond Prometheus - including Datadog, New Relic, PagerDuty, and more. This installation method configures Robusta AI to analyze alerts and incidents from multiple data sources for comprehensive AI-powered investigation and automation capabilities.
 
 .. note::
 
diff --git a/helm/robusta/Chart.lock b/helm/robusta/Chart.lock
@@ -4,6 +4,6 @@ dependencies:
   version: 72.0.1
 - name: holmes
   repository: https://robusta-charts.storage.googleapis.com
-  version: 0.13.3
-digest: sha256:d7053296a08f7f81e9d8b93843f23e73048639c393edb97fbe0b12ba6fbdd118
-generated: "2025-09-03T15:05:39.166399+03:00"
+  version: 0.14.2
+digest: sha256:f9d6fc02a9cd79fad7d694cf2dd9f950632644937f0492e39790ad7f18837cbe
+generated: "2025-09-16T09:58:34.355485+03:00"
diff --git a/helm/robusta/Chart.yaml b/helm/robusta/Chart.yaml
@@ -15,6 +15,6 @@ dependencies:
     condition: enablePrometheusStack
     repository: "https://prometheus-community.github.io/helm-charts"
   - name: holmes
-    version: 0.13.3
+    version: 0.14.2
     condition: enableHolmesGPT
     repository: "https://robusta-charts.storage.googleapis.com"
diff --git a/helm/robusta/charts/holmes-0.13.3.tgz b/helm/robusta/charts/holmes-0.13.3.tgz
diff --git a/helm/robusta/charts/holmes-0.14.2.tgz b/helm/robusta/charts/holmes-0.14.2.tgz
diff --git a/src/robusta/core/model/cluster_status.py b/src/robusta/core/model/cluster_status.py
@@ -23,6 +23,7 @@ class ActivityStats(BaseModel):
     holmesModel: Optional[str]
     clusterTimeZone: str
     errors: List[str]
+    sinksInitializationErrors: bool = False
 
 
 class ClusterStatus(BaseModel):
diff --git a/src/robusta/core/reporting/base.py b/src/robusta/core/reporting/base.py
@@ -17,6 +17,7 @@
 from robusta.core.reporting.consts import FindingSource, FindingSubjectType, FindingType
 from robusta.integrations.kubernetes.api_client_utils import get_namespace_labels
 from robusta.utils.scope import BaseScopeMatcher
+from robusta.utils.time_utils import current_utc_timestamp
 
 
 class BaseBlock(BaseModel):
@@ -289,7 +290,7 @@ def __init__(
         self.fingerprint = (
             fingerprint if fingerprint else self.__calculate_fingerprint(subject, source, aggregation_key)
         )
-        self.starts_at = starts_at if starts_at else datetime.now()
+        self.starts_at = starts_at if starts_at else current_utc_timestamp()
         self.ends_at = ends_at
         self.dirty = False
 
diff --git a/src/robusta/core/reporting/holmes.py b/src/robusta/core/reporting/holmes.py
@@ -73,6 +73,7 @@ class HolmesChatResult(BaseModel):
     files: Optional[List[FileBlock]] = None
     tool_calls: Optional[List[ToolCallResult]] = None
     conversation_history: Optional[List[dict]] = None
+    metadata: Optional[dict] = None
 
 
 class HolmesChatResultsBlock(BaseBlock):
diff --git a/src/robusta/core/sinks/robusta/dal/model_conversion.py b/src/robusta/core/sinks/robusta/dal/model_conversion.py
@@ -26,6 +26,7 @@
 from robusta.core.reporting.holmes import HolmesChatResultsBlock, HolmesResultsBlock, ToolCallResult
 from robusta.core.sinks.transformer import Transformer
 from robusta.utils.parsing import datetime_to_db_str
+from robusta.utils.time_utils import current_utc_timestamp
 
 
 class ModelConversion:
@@ -52,7 +53,7 @@ def to_finding_json(account_id: str, cluster_id: str, finding: Finding):
             "account_id": account_id,
             "video_links": [link.dict() for link in finding.links],  # TD: Migrate column in table.
             "starts_at": datetime_to_db_str(finding.starts_at),
-            "updated_at": datetime_to_db_str(datetime.now()),
+            "updated_at": datetime_to_db_str(current_utc_timestamp()),
         }
 
         if finding.creation_date:
@@ -115,10 +116,13 @@ def append_to_structured_files(files: List[FileBlock], structured_data) -> None:
 
     @staticmethod
     def add_ai_chat_data(structured_data: List[Dict], block: HolmesChatResultsBlock):
+        metadata = block.holmes_result.metadata or {} # type: ignore
+        metadata["type"] = "ai_investigation_result"
+        metadata["createdAt"] = datetime_to_db_str(datetime.now())
         structured_data.append(
             {
                 "type": "markdown",
-                "metadata": {"type": "ai_investigation_result", "createdAt": datetime_to_db_str(datetime.now())},
+                "metadata": metadata,
                 "data": Transformer.to_github_markdown(block.holmes_result.analysis),
             }
         )
diff --git a/src/robusta/core/sinks/robusta/robusta_sink.py b/src/robusta/core/sinks/robusta/robusta_sink.py
@@ -563,6 +563,7 @@ def __update_cluster_status(self):
             holmesModel=holmes_model,
             clusterTimeZone=str(datetime.now().astimezone().tzinfo),
             errors=self.__errors,
+            sinksInitializationErrors=self.registry.has_sink_initialization_errors(),
         )
 
         # checking the status of relay connection
diff --git a/src/robusta/integrations/jira/client.py b/src/robusta/integrations/jira/client.py
@@ -197,7 +197,7 @@ def _resolve_priority(self, priority_name: str) -> dict:
         return {"name": priority_name}
 
     def list_issues(self, search_params: Optional[str] = None):
-        endpoint = "search"
+        endpoint = "search/jql"
         search_params = search_params or ""
         url = self._get_full_jira_url(endpoint)
         query = {"jql": search_params}
diff --git a/src/robusta/model/config.py b/src/robusta/model/config.py
@@ -1,6 +1,6 @@
 import logging
 from collections import defaultdict
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Tuple
 
 from robusta.core.model.env_vars import PROMETHEUS_ENABLED, RUNNER_VERSION
 from robusta.core.playbooks.actions_registry import ActionsRegistry
@@ -46,7 +46,8 @@ def construct_new_sinks(
         new_sinks_config: List[SinkConfigBase],
         existing_sinks: Dict[str, SinkBase],
         registry,
-    ) -> Dict[str, SinkBase]:
+        continue_on_sink_errors: bool = False,
+    ) -> Tuple[Dict[str, SinkBase], bool]:
         new_sink_names = [sink_config.get_name() for sink_config in new_sinks_config]
         # remove deleted sinks
         deleted_sink_names = [sink_name for sink_name in existing_sinks.keys() if sink_name not in new_sink_names]
@@ -56,6 +57,7 @@ def construct_new_sinks(
             del existing_sinks[deleted_sink]
 
         new_sinks: Dict[str, SinkBase] = dict()
+        has_sink_errors = False
         
         # Reload sinks, order does matter and should be loaded & added to the dict by config order.
         for sink_config in new_sinks_config:
@@ -68,24 +70,33 @@ def construct_new_sinks(
 
             sink_name = sink_config.get_name()
             exists_sink = existing_sinks.get(sink_name, None)
-            if not exists_sink:
-                logging.info(f"Adding {type(sink_config)} sink named {sink_name}")
-                new_sinks[sink_name] = SinkFactory.create_sink(sink_config, registry)
-                continue
-
-            is_global_config_changed = exists_sink.is_global_config_changed()
-            is_sink_changed = sink_config.get_params() != exists_sink.params or is_global_config_changed
-            if is_sink_changed:
-                config_change_msg = "due to global config change" if is_global_config_changed else "due to param change"
-                logging.info(f"Updating {type(sink_config)} sink named {sink_config.get_name()} {config_change_msg}")
-                exists_sink.stop()
-                new_sinks[sink_name] = SinkFactory.create_sink(sink_config, registry)
-                continue
-
-            logging.info("Sink %s not changed", sink_name)
-            new_sinks[sink_name] = exists_sink
-
-        return new_sinks
+            
+            try:
+                if not exists_sink:
+                    logging.info(f"Adding {type(sink_config)} sink named {sink_name}")
+                    new_sinks[sink_name] = SinkFactory.create_sink(sink_config, registry)
+                    continue
+
+                is_global_config_changed = exists_sink.is_global_config_changed()
+                is_sink_changed = sink_config.get_params() != exists_sink.params or is_global_config_changed
+                if is_sink_changed:
+                    config_change_msg = "due to global config change" if is_global_config_changed else "due to param change"
+                    logging.info(f"Updating {type(sink_config)} sink named {sink_config.get_name()} {config_change_msg}")
+                    exists_sink.stop()
+                    new_sinks[sink_name] = SinkFactory.create_sink(sink_config, registry)
+                    continue
+
+                logging.info("Sink %s not changed", sink_name)
+                new_sinks[sink_name] = exists_sink
+                
+            except Exception as e:
+                has_sink_errors = True
+                logging.error(f"Failed to initialize sink {sink_name}: {e}", exc_info=True)
+                if not continue_on_sink_errors:
+                    raise
+                # Skip this sink if continue_on_sink_errors is True
+
+        return new_sinks, has_sink_errors
 
 
 class PlaybooksRegistry:
@@ -171,6 +182,7 @@ class Registry:
         prometheus_enabled=PROMETHEUS_ENABLED,
     )
     _pubsub: EventsPubSub = EventsPubSub()
+    _sink_initialization_errors: bool = False
 
     def set_light_actions(self, light_actions: List[str]):
         self._light_actions = light_actions
@@ -231,3 +243,9 @@ def subscribe(self, event_name: str, handler: EventHandler):
 
     def unsubscribe(self, event_name: str, handler: EventHandler):
         self._pubsub.unsubscribe(event_name, handler)
+    
+    def set_sink_initialization_errors(self, has_errors: bool):
+        self._sink_initialization_errors = has_errors
+    
+    def has_sink_initialization_errors(self) -> bool:
+        return self._sink_initialization_errors
diff --git a/src/robusta/runner/config_loader.py b/src/robusta/runner/config_loader.py
@@ -287,10 +287,16 @@ def __prepare_runtime_config(
         sinks_registry: SinksRegistry,
         actions_registry: ActionsRegistry,
         registry: Registry,
-    ) -> (SinksRegistry, PlaybooksRegistry):
+    ) -> tuple[SinksRegistry, PlaybooksRegistry]:
         existing_sinks = sinks_registry.get_all() if sinks_registry else {}
-        new_sinks = SinksRegistry.construct_new_sinks(runner_config.sinks_config, existing_sinks, registry)
+        new_sinks, has_sink_errors = SinksRegistry.construct_new_sinks(
+            runner_config.sinks_config, 
+            existing_sinks, 
+            registry,
+            runner_config.global_config.get("continue_on_sink_errors", False)
+        )
         sinks_registry = SinksRegistry(new_sinks)
+        registry.set_sink_initialization_errors(has_sink_errors)
 
         # TODO we will replace it with a more generic mechanism, as part of the triggers separation task
         # First, we load the internal playbooks, then add the user activated playbooks
diff --git a/src/robusta/utils/time_utils.py b/src/robusta/utils/time_utils.py
@@ -0,0 +1,5 @@
+from datetime import datetime, timezone
+
+
+def current_utc_timestamp() -> datetime:
+    return datetime.now(timezone.utc)
diff --git a/tests/config/test_sink_registry.py b/tests/config/test_sink_registry.py
@@ -24,7 +24,7 @@ def setup_method(self, method):
       self.config_data = yaml.safe_load(CONFIG)
       self.runner_config = RunnerConfig(**self.config_data)
       self.registry = Registry()
-      new_sinks = SinksRegistry.construct_new_sinks(self.runner_config.sinks_config, {}, self.registry)
+      new_sinks, _ = SinksRegistry.construct_new_sinks(self.runner_config.sinks_config, {}, self.registry)
       self.sinks_registry = SinksRegistry(new_sinks)
 
 
@@ -34,20 +34,20 @@ def test_create_sinks(self):
 
   def test_delete_sinks(self):
       del self.runner_config.sinks_config[0]
-      new_sinks = SinksRegistry.construct_new_sinks(self.runner_config.sinks_config, self.sinks_registry.sinks, self.registry)
+      new_sinks, _ = SinksRegistry.construct_new_sinks(self.runner_config.sinks_config, self.sinks_registry.sinks, self.registry)
       assert list(new_sinks.keys()) == ["my_sink2"]
 
       
   def test_add_new_sinks(self):
       self.config_data["sinks_config"].insert(1, {'mail_sink': {'name': 'my_sink3', 'mailto': 'mailtos://user:password@example.com?from=a@x&to=b@y'}})
       self.runner_config = RunnerConfig(**self.config_data)
-      new_sinks = SinksRegistry.construct_new_sinks(self.runner_config.sinks_config, self.sinks_registry.sinks, self.registry)
+      new_sinks, _ = SinksRegistry.construct_new_sinks(self.runner_config.sinks_config, self.sinks_registry.sinks, self.registry)
       assert list(new_sinks.keys()) == ["my_sink1", "my_sink3", "my_sink2"]
 
 
   def test_sink_config_changed(self):
       new_email = "dev@robusta.dev"
       self.runner_config.sinks_config[0].mail_sink.mailto = new_email
-      new_sinks = SinksRegistry.construct_new_sinks(self.runner_config.sinks_config, self.sinks_registry.sinks, self.registry)
+      new_sinks, _ = SinksRegistry.construct_new_sinks(self.runner_config.sinks_config, self.sinks_registry.sinks, self.registry)
       assert new_sinks["my_sink1"].params.mailto == new_email
       assert list(new_sinks.keys()) == ["my_sink1", "my_sink2"]
