Skip to content

Commit 15cdd87

Browse files
Craig ShoemakerCraig Shoemaker
authored andcommitted
add updates
1 parent 092720b commit 15cdd87

1 file changed

Lines changed: 68 additions & 1 deletion

File tree

articles/sre-agent/memory-system.md

Lines changed: 68 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@ description: Use the SRE Agent memory system to build team knowledge that agents
44
author: craigshoemaker
55
ms.author: cshoe
66
ms.reviewer: cshoe
7-
ms.date: 12/18/2025
7+
ms.date: 01/17/2026
88
ms.topic: article
99
ms.service: azure-sre-agent
1010
ms.collection: ce-skilling-ai-copilot
@@ -68,6 +68,30 @@ The `SearchMemory` tool retrieves all memory components. It searches across user
6868
> [!IMPORTANT]
6969
> Don't store secrets, credentials, API keys, or sensitive data in any memory component. Memories are shared across your team and indexed for search.
7070
71+
### Enhanced search parameters
72+
73+
The `SearchNodes` tool supports filtering options for more targeted searches:
74+
75+
| Parameter | Type | Description |
76+
|-----------|------|-------------|
77+
| `entityType` | string | Filter results by entity type (for example, `Incident`, `Service`, `Resource`) |
78+
| `includeNeighbors` | bool | Include connected nodes in the search results |
79+
80+
**Example:**
81+
82+
```text
83+
Search for all incidents related to "database timeout" and include connected resources
84+
```
85+
86+
With `includeNeighbors: true`, the search returns not only matching incident nodes but also their connected:
87+
88+
- Resources
89+
- Services
90+
- Related incidents
91+
- Linked documents
92+
93+
This enables richer context during investigations by showing the full relationship graph around matching nodes.
94+
7195
## Quick start
7296

7397
Begin by establishing foundational knowledge with user memories, and then expand to document storage and automated synchronization as your needs grow.
@@ -205,6 +229,49 @@ The knowledge base provides direct document upload capabilities for runbooks, tr
205229

206230
The portal automatically validates, uploads, and indexes files.
207231

232+
### Upload via agent tool
233+
234+
The agent can upload documents directly to the knowledge base using the `UploadKnowledgeDocument` tool. This is useful when:
235+
236+
- Capturing troubleshooting steps discovered during an investigation
237+
- Adding runbooks generated from incident resolutions
238+
- Programmatically adding documentation without UI access
239+
240+
**Tool: UploadKnowledgeDocument**
241+
242+
| Parameter | Type | Required | Description |
243+
|-----------|------|----------|-------------|
244+
| `fileName` | string | Yes | Filename with extension (for example, `runbook-database-issues.md`). Must be `.md` or `.txt`. |
245+
| `content` | string | Yes | Full document content in plain text or Markdown format. |
246+
| `triggerIndexing` | bool | No | Trigger immediate indexing (default: `true`). Set to `false` for batch uploads. |
247+
248+
**Example usage:**
249+
250+
Ask the agent:
251+
252+
> "Save our troubleshooting steps for the database timeout issue to the knowledge base"
253+
254+
The agent uses `UploadKnowledgeDocument` to:
255+
256+
1. Create a document with appropriate filename
257+
1. Format the content in Markdown
258+
1. Upload to Azure Blob Storage
259+
1. Trigger indexing for immediate searchability
260+
261+
**Constraints:**
262+
263+
- Maximum file size: 16 MB
264+
- Supported extensions: `.md`, `.txt` only
265+
- If a document with the same filename exists, it's overwritten
266+
267+
**Error handling:**
268+
269+
| Error | Resolution |
270+
|-------|------------|
271+
| "Agent memory is disabled" | Enable agent memory in configuration |
272+
| "Invalid file extension" | Use `.md` or `.txt` extension only |
273+
| "Document content exceeds maximum size" | Split large documents into smaller files |
274+
208275
### Manage documents
209276

210277
- **View**: Go to **Settings** > **Knowledge Base** to see all uploaded documents.

0 commit comments

Comments
 (0)