holmesgpt 0.13.0__py3-none-any.whl → 0.13.2__py3-none-any.whl

holmes/plugins/toolsets/datadog/toolset_datadog_logs.py

@@ -1,3 +1,4 @@
+import os
 from enum import Enum
 import json
 import logging
@@ -141,22 +142,25 @@ class DatadogLogsToolset(BasePodLoggingToolset):
 
     @property
     def supported_capabilities(self) -> Set[LoggingCapability]:
-        """Datadog logs API only supports substring matching, no exclude filter"""
-        return set()  # No regex support, no exclude filter
+        """Datadog logs API supports historical data and substring matching"""
+        return {
+            LoggingCapability.HISTORICAL_DATA
+        }  # No regex support, no exclude filter, but supports historical data
 
     def __init__(self):
         super().__init__(
             name="datadog/logs",
-            description="Toolset for interacting with Datadog to fetch logs",
+            description="Toolset for fetching logs from Datadog, including historical data for pods no longer in the cluster",
             docs_url="https://docs.datadoghq.com/api/latest/logs/",
             icon_url="https://imgix.datadoghq.com//img/about/presskit/DDlogo.jpg",
             prerequisites=[CallablePrerequisite(callable=self.prerequisites_callable)],
-            tools=[
-                PodLoggingTool(self),
-            ],
+            tools=[],  # Initialize with empty tools first
             experimental=True,
             tags=[ToolsetTag.CORE],
         )
+        # Now that parent is initialized and self.name exists, create the tool
+        self.tools = [PodLoggingTool(self)]
+        self._reload_instructions()
 
     def logger_name(self) -> str:
         return "DataDog"
@@ -272,3 +276,10 @@ class DatadogLogsToolset(BasePodLoggingToolset):
             "dd_app_key": "your-datadog-application-key",
             "site_api_url": "https://api.datadoghq.com",
         }
+
+    def _reload_instructions(self):
+        """Load Datadog logs specific troubleshooting instructions."""
+        template_file_path = os.path.abspath(
+            os.path.join(os.path.dirname(__file__), "datadog_logs_instructions.jinja2")
+        )
+        self._load_llm_instructions(jinja_template=f"file://{template_file_path}")

holmes/plugins/toolsets/datadog/toolset_datadog_metrics.py

@@ -54,7 +54,7 @@ class ListActiveMetrics(BaseDatadogMetricsTool):
     def __init__(self, toolset: "DatadogMetricsToolset"):
         super().__init__(
             name="list_active_datadog_metrics",
-            description=f"List active metrics from the last {ACTIVE_METRICS_DEFAULT_LOOK_BACK_HOURS} hours. This includes metrics that have actively reported data points.",
+            description=f"List active metrics from Datadog for the last {ACTIVE_METRICS_DEFAULT_LOOK_BACK_HOURS} hours. This includes metrics that have actively reported data points, including from pods no longer in the cluster.",
             parameters={
                 "from_time": ToolParameter(
                     description=f"Start time for listing metrics. Can be an RFC3339 formatted datetime (e.g. '2023-03-01T10:30:00Z') or a negative integer for relative seconds from now (e.g. -86400 for 24 hours ago). Defaults to {ACTIVE_METRICS_DEFAULT_LOOK_BACK_HOURS} hours ago",
@@ -75,7 +75,9 @@ class ListActiveMetrics(BaseDatadogMetricsTool):
             toolset=toolset,
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         if not self.toolset.dd_config:
             return StructuredToolResult(
                 status=ToolResultStatus.ERROR,
@@ -182,7 +184,7 @@ class QueryMetrics(BaseDatadogMetricsTool):
     def __init__(self, toolset: "DatadogMetricsToolset"):
         super().__init__(
             name="query_datadog_metrics",
-            description="Query timeseries data for a specific metric",
+            description="Query timeseries data from Datadog for a specific metric, including historical data for pods no longer in the cluster",
             parameters={
                 "query": ToolParameter(
                     description="The metric query string (e.g., 'system.cpu.user{host:myhost}')",
@@ -215,7 +217,9 @@ class QueryMetrics(BaseDatadogMetricsTool):
             toolset=toolset,
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         if not self.toolset.dd_config:
             return StructuredToolResult(
                 status=ToolResultStatus.ERROR,
@@ -371,7 +375,9 @@ class QueryMetricsMetadata(BaseDatadogMetricsTool):
             toolset=toolset,
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         if not self.toolset.dd_config:
             return StructuredToolResult(
                 status=ToolResultStatus.ERROR,
@@ -485,7 +491,9 @@ class ListMetricTags(BaseDatadogMetricsTool):
             toolset=toolset,
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         if not self.toolset.dd_config:
             return StructuredToolResult(
                 status=ToolResultStatus.ERROR,
@@ -562,7 +570,7 @@ class DatadogMetricsToolset(Toolset):
     def __init__(self):
         super().__init__(
             name="datadog/metrics",
-            description="Toolset for interacting with Datadog to fetch metrics and metadata",
+            description="Toolset for fetching metrics and metadata from Datadog, including historical data for pods no longer in the cluster",
             docs_url="https://docs.datadoghq.com/api/latest/metrics/",
             icon_url="https://imgix.datadoghq.com//img/about/presskit/DDlogo.jpg",
             prerequisites=[CallablePrerequisite(callable=self.prerequisites_callable)],

holmes/plugins/toolsets/datadog/toolset_datadog_rds.py

@@ -92,7 +92,9 @@ class GenerateRDSPerformanceReport(BaseDatadogRDSTool):
             toolset=toolset,
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         if not self.toolset.dd_config:
             return StructuredToolResult(
                 status=ToolResultStatus.ERROR,
@@ -390,7 +392,9 @@ class GetTopWorstPerformingRDSInstances(BaseDatadogRDSTool):
             toolset=toolset,
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         if not self.toolset.dd_config:
             return StructuredToolResult(
                 status=ToolResultStatus.ERROR,

holmes/plugins/toolsets/datadog/toolset_datadog_traces.py

@@ -211,7 +211,9 @@ class FetchDatadogTracesList(BaseDatadogTracesTool):
         filter_str = ", ".join(filters) if filters else "all"
         return f"{toolset_name_for_one_liner(self.toolset.name)}: Fetch Traces ({filter_str})"
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         """Execute the tool to fetch traces."""
         if not self.toolset.dd_config:
             return StructuredToolResult(
@@ -375,7 +377,9 @@ class FetchDatadogTraceById(BaseDatadogTracesTool):
         trace_id = params.get("trace_id", "unknown")
         return f"{toolset_name_for_one_liner(self.toolset.name)}: Fetch Trace Details ({trace_id})"
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         """Execute the tool to fetch trace details."""
         if not self.toolset.dd_config:
             return StructuredToolResult(
@@ -556,7 +560,9 @@ class FetchDatadogSpansByFilter(BaseDatadogTracesTool):
         filter_str = ", ".join(filters) if filters else "all"
         return f"{toolset_name_for_one_liner(self.toolset.name)}: Search Spans ({filter_str})"
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         """Execute the tool to search spans."""
         if not self.toolset.dd_config:
             return StructuredToolResult(

holmes/plugins/toolsets/docker.yaml

@@ -1,7 +1,7 @@
 toolsets:
   docker/core:
     description: "Read access to Docker resources"
-    docs_url: "https://docs.robusta.dev/master/configuration/holmesgpt/toolsets/docker.html"
+    docs_url: "https://holmesgpt.dev/data-sources/builtin-toolsets/docker/"
     icon_url: "https://platform.robusta.dev/demos/docker.svg"
     tags:
       - cli

holmes/plugins/toolsets/git.py

@@ -249,7 +249,9 @@ class GitReadFileWithLineNumbers(Tool):
             toolset=toolset,  # type: ignore
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         filepath = params["filepath"]
         try:
             headers = {"Authorization": f"token {self.toolset.git_credentials}"}
@@ -293,7 +295,9 @@ class GitListFiles(Tool):
             toolset=toolset,  # type: ignore
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         try:
             headers = {"Authorization": f"token {self.toolset.git_credentials}"}
             url = f"https://api.github.com/repos/{self.toolset.git_repo}/git/trees/{self.toolset.git_branch}?recursive=1"
@@ -334,7 +338,9 @@ class GitListOpenPRs(Tool):
             toolset=toolset,  # type: ignore
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         try:
             prs = self.toolset.list_open_prs()
             formatted = [
@@ -402,7 +408,9 @@ class GitExecuteChanges(Tool):
             toolset=toolset,  # type: ignore
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         def error(msg: str) -> StructuredToolResult:
             return StructuredToolResult(
                 status=ToolResultStatus.ERROR,
@@ -620,7 +628,9 @@ class GitUpdatePR(Tool):
             toolset=toolset,  # type: ignore
         )
 
-    def _invoke(self, params: Any) -> StructuredToolResult:
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         try:
             line = params["line"]
             filename = params["filename"]

holmes/plugins/toolsets/grafana/toolset_grafana.py

@@ -1,6 +1,11 @@
 from typing import Dict, List
 from urllib.parse import urlencode, urljoin
-from holmes.core.tools import Tool, ToolParameter
+from holmes.core.tools import (
+    StructuredToolResult,
+    Tool,
+    ToolParameter,
+    ToolResultStatus,
+)
 from holmes.plugins.toolsets.grafana.base_grafana_toolset import BaseGrafanaToolset
 import requests  # type: ignore
 import logging
@@ -38,7 +43,9 @@ class ListAndBuildGrafanaDashboardURLs(Tool):
         )
         self._toolset = toolset
 
-    def _invoke(self, params: Dict) -> str:  # type: ignore
+    def _invoke(
+        self, params: dict, user_approved: bool = False
+    ) -> StructuredToolResult:
         url = urljoin(
             self._toolset._grafana_config.url, "/api/search?query=&type=dash-db"
         )
@@ -82,10 +89,24 @@ class ListAndBuildGrafanaDashboardURLs(Tool):
                     f"Title: {dash['title']}\nURL: {dashboard_url}\n"
                 )
 
-            return "\n".join(formatted_dashboards) or "No dashboards found."
+            return StructuredToolResult(
+                status=ToolResultStatus.SUCCESS
+                if formatted_dashboards
+                else ToolResultStatus.NO_DATA,
+                data="\n".join(formatted_dashboards)
+                if formatted_dashboards
+                else "No dashboards found.",
+                url=url,
+                params=params,
+            )
         except requests.RequestException as e:
             logging.error(f"Error fetching dashboards: {str(e)}")
-            return f"Error fetching dashboards: {str(e)}"
+            return StructuredToolResult(
+                status=ToolResultStatus.ERROR,
+                error=f"Error fetching dashboards: {str(e)}",
+                url=url,
+                params=params,
+            )
 
     def get_parameterized_one_liner(self, params: Dict) -> str:
         return (

holmes/plugins/toolsets/grafana/toolset_grafana_loki.py

@@ -45,12 +45,12 @@ class GrafanaLokiToolset(BasePodLoggingToolset):
             name="grafana/loki",
             description="Fetches kubernetes pods logs from Loki",
             icon_url="https://grafana.com/media/docs/loki/logo-grafana-loki.png",
-            docs_url="https://docs.robusta.dev/master/configuration/holmesgpt/toolsets/grafanaloki.html",
+            docs_url="https://holmesgpt.dev/data-sources/builtin-toolsets/grafanaloki/",
             prerequisites=[CallablePrerequisite(callable=self.prerequisites_callable)],
-            tools=[
-                PodLoggingTool(self),
-            ],
+            tools=[],  # Initialize with empty tools first
         )
+        # Now that parent is initialized and self.name exists, create the tool
+        self.tools = [PodLoggingTool(self)]
 
     def prerequisites_callable(self, config: dict[str, Any]) -> tuple[bool, str]:
         if not config:

holmes/plugins/toolsets/grafana/toolset_grafana_tempo.jinja2

@@ -1,10 +1,12 @@
 Use Tempo when investigating latency or performance issues. Tempo provides traces information for application running on the cluster.
 Assume every application provides tempo traces.
 1. Start by identifying an initial filter to use. This can be a pod name, a deployment name or a service name
-2. Use `fetch_tempo_traces` setting the appropriate query params
+2. Call fetch_tempo_traces_comparative_sample first when investigating performance issues via traces. This tool provides comprehensive analysis for identifying patterns. For other issues not related to performance, you can start with fetch_tempo_traces.
+3. Use `fetch_tempo_traces` setting the appropriate query params
     - Use the min_duration filter to ensure you get traces that trigger the alert when you are investigating a performance issue
     - If possible, use start and end date to narrow down your search.
         - Use fetch_finding_by_id if you are provided with a finding/alert id. It will contain details about when the alert was triggered
     - Use at least one of the following argument to ensure you get relevant traces: `service_name`, `pod_name` or `deployment_name`.
-3. Look at the duration of each span in any single trace and deduct any issues.
-4. ALWAYS fetch the logs for a pod once you identify a span that is taking a long time. There may be an explanation for the slowness in the logs.
+4. When you have a specific trace ID to investigate, use `fetch_tempo_trace_by_id` to get detailed information about that trace.
+5. Look at the duration of each span in any single trace and deduce any issues.
+6. ALWAYS fetch the logs for a pod once you identify a span that is taking a long time. There may be an explanation for the slowness in the logs.