You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: articles/sre-agent/anthropic-sub-processor.md
+102-1Lines changed: 102 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -4,7 +4,7 @@ description: Learn how Anthropic operates as a Microsoft subprocessor in Azure S
4
4
author: craigshoemaker
5
5
ms.author: cshoe
6
6
ms.topic: concept-article
7
-
ms.date: 03/23/2026
7
+
ms.date: 04/07/2026
8
8
ms.service: azure
9
9
ms.custom: references_regions
10
10
---
@@ -72,6 +72,107 @@ When you use Anthropic models in Azure SRE Agent:
72
72
73
73
For more on how Azure SRE Agent handles data, see [Data residency and privacy](data-privacy.md).
74
74
75
+
## Disable Anthropic with Azure Policy
76
+
77
+
You can use Azure Policy to prevent Anthropic from being configured as the AI model provider on your Azure SRE Agent resources.
78
+
79
+
When assigned with the default configuration, this policy blocks Anthropic. Agents that already use Anthropic are flagged as non-compliant but aren't changed automatically.
80
+
81
+
### Create the policy definition
82
+
83
+
1. Save the following policy rule to a file named `deny-anthropic-rules.json`:
1. Save the following parameters to a file named `deny-anthropic-params.json`:
106
+
107
+
```json
108
+
{
109
+
"effect": {
110
+
"type": "String",
111
+
"metadata": {
112
+
"displayName": "Effect",
113
+
"description": "Audit logs non-compliant agents without blocking. Deny prevents agents from being configured with a disallowed provider. Disabled turns off the policy."
114
+
},
115
+
"allowedValues": ["Audit", "Deny", "Disabled"],
116
+
"defaultValue": "Deny"
117
+
},
118
+
"disallowedProviders": {
119
+
"type": "Array",
120
+
"metadata": {
121
+
"displayName": "Disallowed AI model providers",
122
+
"description": "AI model provider values to deny on Azure SRE Agent resources."
123
+
},
124
+
"defaultValue": ["Anthropic"]
125
+
}
126
+
}
127
+
```
128
+
129
+
1. Create the policy definition:
130
+
131
+
```azurecli
132
+
az policy definition create \
133
+
--name "deny-anthropic-provider" \
134
+
--display-name "Azure SRE Agent: Restrict AI model providers" \
135
+
--description "Prevents Azure SRE Agent resources from using Anthropic as the AI model provider." \
136
+
--mode All \
137
+
--rules deny-anthropic-rules.json \
138
+
--params deny-anthropic-params.json
139
+
```
140
+
141
+
### Assign the policy
142
+
143
+
Assign the policy to your subscription:
144
+
145
+
```azurecli
146
+
az policy assignment create \
147
+
--name "deny-anthropic-provider" \
148
+
--display-name "Deny Anthropic on SRE Agent resources" \
149
+
--policy "deny-anthropic-provider" \
150
+
--scope "/subscriptions/<YOUR_SUBSCRIPTION_ID>"
151
+
```
152
+
153
+
To apply the policy across multiple subscriptions, assign it at the management group scope:
154
+
155
+
```azurecli
156
+
az policy assignment create \
157
+
--name "deny-anthropic-provider" \
158
+
--display-name "Deny Anthropic on SRE Agent resources" \
After assignment, any attempt to configure an agent with Anthropic as the provider is blocked. Check compliance status:
166
+
167
+
```azurecli
168
+
az policy state summarize --policy-assignment "deny-anthropic-provider"
169
+
```
170
+
171
+
Existing agents that already use Anthropic are flagged as non-compliant but aren't changed automatically. Update those agents to use Azure OpenAI through the Azure SRE Agent portal under **Settings** > **AI Model Provider**.
172
+
173
+
> [!NOTE]
174
+
> To monitor before enforcing, change the default value for `effect` from `"Deny"` to `"Audit"` in the parameters file. Audit mode logs non-compliant agents without blocking changes.
175
+
75
176
## Related content
76
177
77
178
-[Data residency and privacy in Azure SRE Agent](data-privacy.md)
0 commit comments