@aws/ml-container-creator 0.9.1 → 0.10.3

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (90) hide show
  1. package/LICENSE-THIRD-PARTY +9304 -0
  2. package/bin/cli.js +2 -0
  3. package/config/bootstrap-e2e-stack.json +341 -0
  4. package/config/bootstrap-stack.json +40 -3
  5. package/config/parameter-schema-v2.json +2049 -0
  6. package/config/tune-catalog.json +1781 -0
  7. package/infra/ci-harness/buildspec.yml +1 -0
  8. package/infra/ci-harness/lambda/path-prover/brain.ts +306 -0
  9. package/infra/ci-harness/lambda/path-prover/write-results.ts +152 -0
  10. package/infra/ci-harness/lib/ci-harness-stack.ts +837 -7
  11. package/infra/ci-harness/state-machines/path-prover.asl.json +496 -0
  12. package/package.json +53 -68
  13. package/servers/base-image-picker/index.js +121 -121
  14. package/servers/e2e-status/index.js +297 -0
  15. package/servers/e2e-status/manifest.json +14 -0
  16. package/servers/e2e-status/package.json +15 -0
  17. package/servers/endpoint-picker/LICENSE +202 -0
  18. package/servers/endpoint-picker/index.js +536 -0
  19. package/servers/endpoint-picker/manifest.json +14 -0
  20. package/servers/endpoint-picker/package.json +18 -0
  21. package/servers/hyperpod-cluster-picker/index.js +125 -125
  22. package/servers/instance-sizer/index.js +138 -138
  23. package/servers/instance-sizer/lib/instance-ranker.js +76 -76
  24. package/servers/instance-sizer/lib/model-resolver.js +61 -61
  25. package/servers/instance-sizer/lib/quota-resolver.js +113 -113
  26. package/servers/instance-sizer/lib/vram-estimator.js +31 -31
  27. package/servers/lib/bedrock-client.js +38 -38
  28. package/servers/lib/catalogs/jumpstart-public.json +101 -16
  29. package/servers/lib/catalogs/model-servers.json +201 -3
  30. package/servers/lib/catalogs/models.json +182 -26
  31. package/servers/lib/custom-validators.js +13 -13
  32. package/servers/lib/dynamic-resolver.js +4 -4
  33. package/servers/marketplace-picker/index.js +342 -0
  34. package/servers/marketplace-picker/manifest.json +14 -0
  35. package/servers/marketplace-picker/package.json +18 -0
  36. package/servers/model-picker/index.js +382 -382
  37. package/servers/region-picker/index.js +56 -56
  38. package/servers/workload-picker/LICENSE +202 -0
  39. package/servers/workload-picker/catalogs/workload-profiles.json +67 -0
  40. package/servers/workload-picker/index.js +171 -0
  41. package/servers/workload-picker/manifest.json +16 -0
  42. package/servers/workload-picker/package.json +16 -0
  43. package/src/app.js +4 -390
  44. package/src/lib/bootstrap-command-handler.js +710 -1148
  45. package/src/lib/bootstrap-config.js +36 -0
  46. package/src/lib/bootstrap-profile-manager.js +641 -0
  47. package/src/lib/bootstrap-provisioners.js +421 -0
  48. package/src/lib/ci-register-helpers.js +74 -0
  49. package/src/lib/config-loader.js +408 -0
  50. package/src/lib/config-manager.js +66 -1685
  51. package/src/lib/config-mcp-client.js +118 -0
  52. package/src/lib/config-validator.js +634 -0
  53. package/src/lib/cuda-resolver.js +149 -0
  54. package/src/lib/e2e-catalog-validator.js +251 -3
  55. package/src/lib/e2e-ci-recorder.js +103 -0
  56. package/src/lib/generated/cli-options.js +315 -311
  57. package/src/lib/generated/parameter-matrix.js +671 -0
  58. package/src/lib/generated/validation-rules.js +71 -71
  59. package/src/lib/marketplace-flow.js +276 -0
  60. package/src/lib/mcp-query-runner.js +768 -0
  61. package/src/lib/parameter-schema-validator.js +62 -18
  62. package/src/lib/path-prover-brain.js +607 -0
  63. package/src/lib/prompt-runner.js +41 -1504
  64. package/src/lib/prompts/feature-prompts.js +172 -0
  65. package/src/lib/prompts/index.js +48 -0
  66. package/src/lib/prompts/infrastructure-prompts.js +690 -0
  67. package/src/lib/prompts/model-prompts.js +552 -0
  68. package/src/lib/prompts/project-prompts.js +82 -0
  69. package/src/lib/prompts.js +2 -1446
  70. package/src/lib/registry-command-handler.js +135 -3
  71. package/src/lib/secrets-prompt-runner.js +251 -0
  72. package/src/lib/template-variable-resolver.js +422 -0
  73. package/src/lib/tune-catalog-validator.js +37 -4
  74. package/templates/Dockerfile +9 -0
  75. package/templates/code/adapter_sidecar.py +444 -0
  76. package/templates/code/serve +6 -0
  77. package/templates/code/serve.d/vllm.ejs +1 -1
  78. package/templates/do/.benchmark_writer.py +1476 -0
  79. package/templates/do/.tune_helper.py +982 -57
  80. package/templates/do/__pycache__/.benchmark_writer.cpython-312.pyc +0 -0
  81. package/templates/do/adapter +149 -0
  82. package/templates/do/benchmark +639 -85
  83. package/templates/do/config +108 -5
  84. package/templates/do/deploy.d/managed-inference.ejs +192 -11
  85. package/templates/do/optimize +106 -37
  86. package/templates/do/register +89 -0
  87. package/templates/do/test +13 -0
  88. package/templates/do/tune +378 -59
  89. package/templates/do/validate +44 -4
  90. package/config/parameter-schema.json +0 -88
@@ -23,18 +23,18 @@
23
23
  * BEDROCK_REGION - AWS region for Bedrock API calls (fallback: AWS_REGION, then us-east-1)
24
24
  */
25
25
 
26
- import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js'
27
- import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js'
28
- import { z } from 'zod'
29
- import { readFileSync } from 'node:fs'
30
- import { fileURLToPath } from 'node:url'
31
- import { resolve, dirname } from 'node:path'
32
- import { queryBedrock } from '../lib/bedrock-client.js'
26
+ import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
27
+ import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
28
+ import { z } from 'zod';
29
+ import { readFileSync } from 'node:fs';
30
+ import { fileURLToPath } from 'node:url';
31
+ import { resolve, dirname } from 'node:path';
32
+ import { queryBedrock } from '../lib/bedrock-client.js';
33
33
 
34
34
  // ── Catalog loader ───────────────────────────────────────────────────────────
35
35
 
36
- const __filename = fileURLToPath(import.meta.url)
37
- const __dirname = dirname(__filename)
36
+ const __filename = fileURLToPath(import.meta.url);
37
+ const __dirname = dirname(__filename);
38
38
 
39
39
  /**
40
40
  * Load and parse a JSON catalog file relative to the server directory.
@@ -44,37 +44,37 @@ const __dirname = dirname(__filename)
44
44
  * @returns {any} Parsed JSON content
45
45
  */
46
46
  function loadCatalog(relativePath) {
47
- const fullPath = resolve(__dirname, relativePath)
48
- let raw
47
+ const fullPath = resolve(__dirname, relativePath);
48
+ let raw;
49
49
  try {
50
- raw = readFileSync(fullPath, 'utf8')
50
+ raw = readFileSync(fullPath, 'utf8');
51
51
  } catch (err) {
52
- throw new Error(`Catalog file not found: ${fullPath}`)
52
+ throw new Error(`Catalog file not found: ${fullPath}`);
53
53
  }
54
54
  try {
55
- return JSON.parse(raw)
55
+ return JSON.parse(raw);
56
56
  } catch (err) {
57
- throw new Error(`Failed to parse catalog ${fullPath}: ${err.message}`)
57
+ throw new Error(`Failed to parse catalog ${fullPath}: ${err.message}`);
58
58
  }
59
59
  }
60
60
 
61
61
  // ── Load catalogs from JSON files ─────────────────────────────────────────────
62
62
 
63
- let AWS_REGIONS
64
- let VALID_REGION_CODES
63
+ let AWS_REGIONS;
64
+ let VALID_REGION_CODES;
65
65
 
66
66
  try {
67
- AWS_REGIONS = loadCatalog('../lib/catalogs/regions.json')
68
- VALID_REGION_CODES = new Set(AWS_REGIONS.map(r => r.code))
67
+ AWS_REGIONS = loadCatalog('../lib/catalogs/regions.json');
68
+ VALID_REGION_CODES = new Set(AWS_REGIONS.map(r => r.code));
69
69
  } catch (err) {
70
- process.stderr.write(`[region-picker] Fatal: ${err.message}\n`)
71
- process.exit(1)
70
+ process.stderr.write(`[region-picker] Fatal: ${err.message}\n`);
71
+ process.exit(1);
72
72
  }
73
73
 
74
74
  // Bedrock / smart-mode configuration
75
- const SMART_MODE = process.env.BEDROCK_SMART === 'true'
76
- const BEDROCK_MODEL = process.env.BEDROCK_MODEL || 'global.anthropic.claude-sonnet-4-20250514-v1:0'
77
- const BEDROCK_REGION = process.env.BEDROCK_REGION || process.env.AWS_REGION || 'us-east-1'
75
+ const SMART_MODE = process.env.BEDROCK_SMART === 'true';
76
+ const BEDROCK_MODEL = process.env.BEDROCK_MODEL || 'global.anthropic.claude-sonnet-4-20250514-v1:0';
77
+ const BEDROCK_REGION = process.env.BEDROCK_REGION || process.env.AWS_REGION || 'us-east-1';
78
78
 
79
79
  /**
80
80
  * Per-server configuration passed to the shared Bedrock client.
@@ -105,7 +105,7 @@ Rules:
105
105
  maxTokens: 1024,
106
106
  modelId: BEDROCK_MODEL,
107
107
  region: BEDROCK_REGION
108
- }
108
+ };
109
109
 
110
110
  /**
111
111
  * Filter AWS_REGIONS by a case-insensitive substring match against
@@ -116,42 +116,42 @@ Rules:
116
116
  * @returns {{ values: object, choices: object }}
117
117
  */
118
118
  function filterRegions(searchTerm, limit) {
119
- let matched
119
+ let matched;
120
120
 
121
121
  if (searchTerm) {
122
- const term = searchTerm.toLowerCase()
122
+ const term = searchTerm.toLowerCase();
123
123
  matched = AWS_REGIONS.filter(
124
124
  r => r.code.toLowerCase().includes(term) ||
125
125
  r.labels.some(l => l.toLowerCase().includes(term))
126
- )
126
+ );
127
127
  } else {
128
- matched = AWS_REGIONS
128
+ matched = AWS_REGIONS;
129
129
  }
130
130
 
131
- const codes = matched.map(r => r.code).slice(0, limit)
131
+ const codes = matched.map(r => r.code).slice(0, limit);
132
132
 
133
133
  if (codes.length === 0) {
134
- return { values: {}, choices: { awsRegion: [] } }
134
+ return { values: {}, choices: { awsRegion: [] } };
135
135
  }
136
136
 
137
137
  return {
138
138
  values: { awsRegion: codes[0] },
139
139
  choices: { awsRegion: codes }
140
- }
140
+ };
141
141
  }
142
142
 
143
143
  /**
144
144
  * Log to stderr so it doesn't interfere with MCP stdio protocol on stdout.
145
145
  */
146
146
  function log(message) {
147
- process.stderr.write(`[region-picker] ${message}\n`)
147
+ process.stderr.write(`[region-picker] ${message}\n`);
148
148
  }
149
149
 
150
150
  // Create MCP server
151
151
  const server = new McpServer({
152
152
  name: 'region-picker',
153
153
  version: '1.0.0'
154
- })
154
+ });
155
155
 
156
156
  // Register the get_regions tool
157
157
  server.tool(
@@ -170,37 +170,37 @@ server.tool(
170
170
  type: 'text',
171
171
  text: JSON.stringify({ values: {}, choices: {} })
172
172
  }]
173
- }
173
+ };
174
174
  }
175
175
 
176
- const searchTerm = context?.regionSearch
177
- let result
176
+ const searchTerm = context?.regionSearch;
177
+ let result;
178
178
 
179
179
  // Smart mode: try Bedrock first
180
180
  if (SMART_MODE) {
181
- log('[smart] Smart mode enabled, querying Amazon Bedrock...')
182
- const bedrockResult = await queryBedrock(SERVER_CONFIG, parameters, limit, context || {})
181
+ log('[smart] Smart mode enabled, querying Amazon Bedrock...');
182
+ const bedrockResult = await queryBedrock(SERVER_CONFIG, parameters, limit, context || {});
183
183
 
184
184
  if (bedrockResult?.values?.awsRegion && VALID_REGION_CODES.has(bedrockResult.values.awsRegion)) {
185
- const bedrockValue = bedrockResult.values.awsRegion
186
- log(`[smart] Using Bedrock recommendation: ${bedrockValue}`)
185
+ const bedrockValue = bedrockResult.values.awsRegion;
186
+ log(`[smart] Using Bedrock recommendation: ${bedrockValue}`);
187
187
 
188
188
  // Pad with static results, deduplicating the Bedrock pick
189
- const staticResult = filterRegions(searchTerm, limit)
190
- const staticCodes = staticResult.choices.awsRegion || []
191
- const combined = [bedrockValue, ...staticCodes.filter(c => c !== bedrockValue)]
189
+ const staticResult = filterRegions(searchTerm, limit);
190
+ const staticCodes = staticResult.choices.awsRegion || [];
191
+ const combined = [bedrockValue, ...staticCodes.filter(c => c !== bedrockValue)];
192
192
 
193
193
  result = {
194
194
  values: { awsRegion: bedrockValue },
195
195
  choices: { awsRegion: combined.slice(0, limit) }
196
- }
196
+ };
197
197
  } else {
198
- log('[smart] Bedrock did not return usable results, falling back to static filtering')
199
- result = filterRegions(searchTerm, limit)
198
+ log('[smart] Bedrock did not return usable results, falling back to static filtering');
199
+ result = filterRegions(searchTerm, limit);
200
200
  }
201
201
  } else {
202
202
  // Static mode (default)
203
- result = filterRegions(searchTerm, limit)
203
+ result = filterRegions(searchTerm, limit);
204
204
  }
205
205
 
206
206
  return {
@@ -208,23 +208,23 @@ server.tool(
208
208
  type: 'text',
209
209
  text: JSON.stringify(result)
210
210
  }]
211
- }
211
+ };
212
212
  }
213
- )
213
+ );
214
214
 
215
215
  // Export for standalone testing
216
- export { loadCatalog, filterRegions, AWS_REGIONS, VALID_REGION_CODES }
216
+ export { loadCatalog, filterRegions, AWS_REGIONS, VALID_REGION_CODES };
217
217
 
218
218
  // Guard MCP transport — only connect when run as main module
219
- const isMain = process.argv[1] && resolve(process.argv[1]) === __filename
219
+ const isMain = process.argv[1] && resolve(process.argv[1]) === __filename;
220
220
 
221
221
  if (isMain) {
222
222
  if (SMART_MODE) {
223
- log(`Smart mode enabled (model: ${BEDROCK_MODEL}, region: ${BEDROCK_REGION})`)
223
+ log(`Smart mode enabled (model: ${BEDROCK_MODEL}, region: ${BEDROCK_REGION})`);
224
224
  } else {
225
- log('Static mode (set BEDROCK_SMART=true to enable Bedrock-powered recommendations)')
225
+ log('Static mode (set BEDROCK_SMART=true to enable Bedrock-powered recommendations)');
226
226
  }
227
227
 
228
- const transport = new StdioServerTransport()
229
- await server.connect(transport)
228
+ const transport = new StdioServerTransport();
229
+ await server.connect(transport);
230
230
  }
@@ -0,0 +1,202 @@
1
+
2
+ Apache License
3
+ Version 2.0, January 2004
4
+ http://www.apache.org/licenses/
5
+
6
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
7
+
8
+ 1. Definitions.
9
+
10
+ "License" shall mean the terms and conditions for use, reproduction,
11
+ and distribution as defined by Sections 1 through 9 of this document.
12
+
13
+ "Licensor" shall mean the copyright owner or entity authorized by
14
+ the copyright owner that is granting the License.
15
+
16
+ "Legal Entity" shall mean the union of the acting entity and all
17
+ other entities that control, are controlled by, or are under common
18
+ control with that entity. For the purposes of this definition,
19
+ "control" means (i) the power, direct or indirect, to cause the
20
+ direction or management of such entity, whether by contract or
21
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
22
+ outstanding shares, or (iii) beneficial ownership of such entity.
23
+
24
+ "You" (or "Your") shall mean an individual or Legal Entity
25
+ exercising permissions granted by this License.
26
+
27
+ "Source" form shall mean the preferred form for making modifications,
28
+ including but not limited to software source code, documentation
29
+ source, and configuration files.
30
+
31
+ "Object" form shall mean any form resulting from mechanical
32
+ transformation or translation of a Source form, including but
33
+ not limited to compiled object code, generated documentation,
34
+ and conversions to other media types.
35
+
36
+ "Work" shall mean the work of authorship, whether in Source or
37
+ Object form, made available under the License, as indicated by a
38
+ copyright notice that is included in or attached to the work
39
+ (an example is provided in the Appendix below).
40
+
41
+ "Derivative Works" shall mean any work, whether in Source or Object
42
+ form, that is based on (or derived from) the Work and for which the
43
+ editorial revisions, annotations, elaborations, or other modifications
44
+ represent, as a whole, an original work of authorship. For the purposes
45
+ of this License, Derivative Works shall not include works that remain
46
+ separable from, or merely link (or bind by name) to the interfaces of,
47
+ the Work and Derivative Works thereof.
48
+
49
+ "Contribution" shall mean any work of authorship, including
50
+ the original version of the Work and any modifications or additions
51
+ to that Work or Derivative Works thereof, that is intentionally
52
+ submitted to Licensor for inclusion in the Work by the copyright owner
53
+ or by an individual or Legal Entity authorized to submit on behalf of
54
+ the copyright owner. For the purposes of this definition, "submitted"
55
+ means any form of electronic, verbal, or written communication sent
56
+ to the Licensor or its representatives, including but not limited to
57
+ communication on electronic mailing lists, source code control systems,
58
+ and issue tracking systems that are managed by, or on behalf of, the
59
+ Licensor for the purpose of discussing and improving the Work, but
60
+ excluding communication that is conspicuously marked or otherwise
61
+ designated in writing by the copyright owner as "Not a Contribution."
62
+
63
+ "Contributor" shall mean Licensor and any individual or Legal Entity
64
+ on behalf of whom a Contribution has been received by Licensor and
65
+ subsequently incorporated within the Work.
66
+
67
+ 2. Grant of Copyright License. Subject to the terms and conditions of
68
+ this License, each Contributor hereby grants to You a perpetual,
69
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
70
+ copyright license to reproduce, prepare Derivative Works of,
71
+ publicly display, publicly perform, sublicense, and distribute the
72
+ Work and such Derivative Works in Source or Object form.
73
+
74
+ 3. Grant of Patent License. Subject to the terms and conditions of
75
+ this License, each Contributor hereby grants to You a perpetual,
76
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
77
+ (except as stated in this section) patent license to make, have made,
78
+ use, offer to sell, sell, import, and otherwise transfer the Work,
79
+ where such license applies only to those patent claims licensable
80
+ by such Contributor that are necessarily infringed by their
81
+ Contribution(s) alone or by combination of their Contribution(s)
82
+ with the Work to which such Contribution(s) was submitted. If You
83
+ institute patent litigation against any entity (including a
84
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
85
+ or a Contribution incorporated within the Work constitutes direct
86
+ or contributory patent infringement, then any patent licenses
87
+ granted to You under this License for that Work shall terminate
88
+ as of the date such litigation is filed.
89
+
90
+ 4. Redistribution. You may reproduce and distribute copies of the
91
+ Work or Derivative Works thereof in any medium, with or without
92
+ modifications, and in Source or Object form, provided that You
93
+ meet the following conditions:
94
+
95
+ (a) You must give any other recipients of the Work or
96
+ Derivative Works a copy of this License; and
97
+
98
+ (b) You must cause any modified files to carry prominent notices
99
+ stating that You changed the files; and
100
+
101
+ (c) You must retain, in the Source form of any Derivative Works
102
+ that You distribute, all copyright, patent, trademark, and
103
+ attribution notices from the Source form of the Work,
104
+ excluding those notices that do not pertain to any part of
105
+ the Derivative Works; and
106
+
107
+ (d) If the Work includes a "NOTICE" text file as part of its
108
+ distribution, then any Derivative Works that You distribute must
109
+ include a readable copy of the attribution notices contained
110
+ within such NOTICE file, excluding those notices that do not
111
+ pertain to any part of the Derivative Works, in at least one
112
+ of the following places: within a NOTICE text file distributed
113
+ as part of the Derivative Works; within the Source form or
114
+ documentation, if provided along with the Derivative Works; or,
115
+ within a display generated by the Derivative Works, if and
116
+ wherever such third-party notices normally appear. The contents
117
+ of the NOTICE file are for informational purposes only and
118
+ do not modify the License. You may add Your own attribution
119
+ notices within Derivative Works that You distribute, alongside
120
+ or as an addendum to the NOTICE text from the Work, provided
121
+ that such additional attribution notices cannot be construed
122
+ as modifying the License.
123
+
124
+ You may add Your own copyright statement to Your modifications and
125
+ may provide additional or different license terms and conditions
126
+ for use, reproduction, or distribution of Your modifications, or
127
+ for any such Derivative Works as a whole, provided Your use,
128
+ reproduction, and distribution of the Work otherwise complies with
129
+ the conditions stated in this License.
130
+
131
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
132
+ any Contribution intentionally submitted for inclusion in the Work
133
+ by You to the Licensor shall be under the terms and conditions of
134
+ this License, without any additional terms or conditions.
135
+ Notwithstanding the above, nothing herein shall supersede or modify
136
+ the terms of any separate license agreement you may have executed
137
+ with Licensor regarding such Contributions.
138
+
139
+ 6. Trademarks. This License does not grant permission to use the trade
140
+ names, trademarks, service marks, or product names of the Licensor,
141
+ except as required for reasonable and customary use in describing the
142
+ origin of the Work and reproducing the content of the NOTICE file.
143
+
144
+ 7. Disclaimer of Warranty. Unless required by applicable law or
145
+ agreed to in writing, Licensor provides the Work (and each
146
+ Contributor provides its Contributions) on an "AS IS" BASIS,
147
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
148
+ implied, including, without limitation, any warranties or conditions
149
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
150
+ PARTICULAR PURPOSE. You are solely responsible for determining the
151
+ appropriateness of using or redistributing the Work and assume any
152
+ risks associated with Your exercise of permissions under this License.
153
+
154
+ 8. Limitation of Liability. In no event and under no legal theory,
155
+ whether in tort (including negligence), contract, or otherwise,
156
+ unless required by applicable law (such as deliberate and grossly
157
+ negligent acts) or agreed to in writing, shall any Contributor be
158
+ liable to You for damages, including any direct, indirect, special,
159
+ incidental, or consequential damages of any character arising as a
160
+ result of this License or out of the use or inability to use the
161
+ Work (including but not limited to damages for loss of goodwill,
162
+ work stoppage, computer failure or malfunction, or any and all
163
+ other commercial damages or losses), even if such Contributor
164
+ has been advised of the possibility of such damages.
165
+
166
+ 9. Accepting Warranty or Additional Liability. While redistributing
167
+ the Work or Derivative Works thereof, You may choose to offer,
168
+ and charge a fee for, acceptance of support, warranty, indemnity,
169
+ or other liability obligations and/or rights consistent with this
170
+ License. However, in accepting such obligations, You may act only
171
+ on Your own behalf and on Your sole responsibility, not on behalf
172
+ of any other Contributor, and only if You agree to indemnify,
173
+ defend, and hold each Contributor harmless for any liability
174
+ incurred by, or claims asserted against, such Contributor by reason
175
+ of your accepting any such warranty or additional liability.
176
+
177
+ END OF TERMS AND CONDITIONS
178
+
179
+ APPENDIX: How to apply the Apache License to your work.
180
+
181
+ To apply the Apache License to your work, attach the following
182
+ boilerplate notice, with the fields enclosed by brackets "[]"
183
+ replaced with your own identifying information. (Don't include
184
+ the brackets!) The text should be enclosed in the appropriate
185
+ comment syntax for the file format. We also recommend that a
186
+ file or class name and description of purpose be included on the
187
+ same "printed page" as the copyright notice for easier
188
+ identification within third-party archives.
189
+
190
+ Copyright [yyyy] [name of copyright owner]
191
+
192
+ Licensed under the Apache License, Version 2.0 (the "License");
193
+ you may not use this file except in compliance with the License.
194
+ You may obtain a copy of the License at
195
+
196
+ http://www.apache.org/licenses/LICENSE-2.0
197
+
198
+ Unless required by applicable law or agreed to in writing, software
199
+ distributed under the License is distributed on an "AS IS" BASIS,
200
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
201
+ See the License for the specific language governing permissions and
202
+ limitations under the License.
@@ -0,0 +1,67 @@
1
+ {
2
+ "workloads": {
3
+ "sample": {
4
+ "description": "POC Sample Workload",
5
+ "input_tokens_mean": 100,
6
+ "output_tokens_mean": 100,
7
+ "streaming": true,
8
+ "dataset": "synthetic",
9
+ "concurrency_levels": [2],
10
+ "use_case": "POC Sample Workload"
11
+ },
12
+ "multi_turn_chat": {
13
+ "description": "Copilots, assistants, customer support — short exchanges with conversation history",
14
+ "input_tokens_mean": 550,
15
+ "output_tokens_mean": 150,
16
+ "streaming": true,
17
+ "dataset": "synthetic",
18
+ "concurrency_levels": [1, 4, 8, 16],
19
+ "use_case": "Interactive chat, low latency"
20
+ },
21
+ "rag_document_qa": {
22
+ "description": "Knowledge bases, legal & medical search — long context retrieval with focused answers",
23
+ "input_tokens_mean": 2000,
24
+ "output_tokens_mean": 500,
25
+ "streaming": true,
26
+ "dataset": "synthetic",
27
+ "concurrency_levels": [1, 4, 8],
28
+ "use_case": "RAG, document Q&A"
29
+ },
30
+ "agent_tool_calling": {
31
+ "description": "Coding agents, research agents, automation — short structured outputs (JSON)",
32
+ "input_tokens_mean": 200,
33
+ "output_tokens_mean": 100,
34
+ "streaming": false,
35
+ "dataset": "synthetic",
36
+ "concurrency_levels": [1, 4, 8, 16, 32],
37
+ "use_case": "Agents, structured output"
38
+ },
39
+ "long_context_scaling": {
40
+ "description": "Prefill stress test — very long inputs with moderate outputs",
41
+ "input_tokens_mean": 8000,
42
+ "output_tokens_mean": 1000,
43
+ "streaming": true,
44
+ "dataset": "synthetic",
45
+ "concurrency_levels": [1, 2, 4],
46
+ "use_case": "Long context, summarization"
47
+ },
48
+ "production_traffic_mix": {
49
+ "description": "Aggregate real traffic simulation — mixed input/output lengths",
50
+ "input_tokens_mean": 1000,
51
+ "output_tokens_mean": 300,
52
+ "streaming": true,
53
+ "dataset": "synthetic",
54
+ "concurrency_levels": [4, 8, 16, 32],
55
+ "use_case": "Production fleet sizing"
56
+ },
57
+ "shared_system_prompt": {
58
+ "description": "Prefix caching / prompt reuse — long shared prefix with short user queries",
59
+ "input_tokens_mean": 1000,
60
+ "output_tokens_mean": 200,
61
+ "streaming": true,
62
+ "dataset": "synthetic",
63
+ "concurrency_levels": [4, 8, 16, 32],
64
+ "use_case": "Prefix caching validation"
65
+ }
66
+ }
67
+ }
@@ -0,0 +1,171 @@
1
+ #!/usr/bin/env node
2
+ // Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
3
+ // SPDX-License-Identifier: Apache-2.0
4
+
5
+ /**
6
+ * Workload Picker MCP Server
7
+ *
8
+ * A bundled MCP server that provides named benchmark workload profiles
9
+ * for use with `do/benchmark`. Workloads define token distributions,
10
+ * concurrency levels, streaming mode, and intended use cases.
11
+ *
12
+ * Tools:
13
+ * - list_workloads: Returns all available workload names with descriptions
14
+ * - get_workload_profile: Returns full parameters for a named workload
15
+ *
16
+ * The catalog is loaded from ./catalogs/workload-profiles.json and can be
17
+ * extended by adding entries to that file.
18
+ */
19
+
20
+ import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
21
+ import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
22
+ import { z } from 'zod';
23
+ import { readFileSync } from 'node:fs';
24
+ import { fileURLToPath } from 'node:url';
25
+ import { resolve, dirname } from 'node:path';
26
+
27
+ // ── Catalog loader ───────────────────────────────────────────────────────────
28
+
29
+ const __filename = fileURLToPath(import.meta.url);
30
+ const __dirname = dirname(__filename);
31
+
32
+ /**
33
+ * Load and parse a JSON catalog file relative to the server directory.
34
+ * Throws on missing file or invalid JSON with the file path in the message.
35
+ *
36
+ * @param {string} relativePath - Path relative to server dir
37
+ * @returns {any} Parsed JSON content
38
+ */
39
+ function loadCatalog(relativePath) {
40
+ const fullPath = resolve(__dirname, relativePath);
41
+ let raw;
42
+ try {
43
+ raw = readFileSync(fullPath, 'utf8');
44
+ } catch (err) {
45
+ throw new Error(`Catalog file not found: ${fullPath}`);
46
+ }
47
+ try {
48
+ return JSON.parse(raw);
49
+ } catch (err) {
50
+ throw new Error(`Failed to parse catalog ${fullPath}: ${err.message}`);
51
+ }
52
+ }
53
+
54
+ // ── Load workload profiles ───────────────────────────────────────────────────
55
+
56
+ let WORKLOAD_CATALOG;
57
+
58
+ try {
59
+ WORKLOAD_CATALOG = loadCatalog('./catalogs/workload-profiles.json');
60
+ } catch (err) {
61
+ process.stderr.write(`[workload-picker] Fatal: ${err.message}\n`);
62
+ process.exit(1);
63
+ }
64
+
65
+ const WORKLOAD_NAMES = Object.keys(WORKLOAD_CATALOG.workloads);
66
+
67
+ // ── Helper functions ─────────────────────────────────────────────────────────
68
+
69
+ /**
70
+ * List all available workloads with name + description + use_case.
71
+ *
72
+ * @returns {{ workloads: Array<{ name: string, description: string, use_case: string }> }}
73
+ */
74
+ export function listWorkloads() {
75
+ const workloads = WORKLOAD_NAMES.map(name => ({
76
+ name,
77
+ description: WORKLOAD_CATALOG.workloads[name].description,
78
+ use_case: WORKLOAD_CATALOG.workloads[name].use_case
79
+ }));
80
+ return { workloads };
81
+ }
82
+
83
+ /**
84
+ * Get full workload profile by name.
85
+ *
86
+ * @param {string} workloadName - One of the defined workload names
87
+ * @returns {object|null} Full workload profile or null if not found
88
+ */
89
+ export function getWorkloadProfile(workloadName) {
90
+ const profile = WORKLOAD_CATALOG.workloads[workloadName];
91
+ if (!profile) return null;
92
+ return { name: workloadName, ...profile };
93
+ }
94
+
95
+ /**
96
+ * Log to stderr so it doesn't interfere with MCP stdio protocol on stdout.
97
+ */
98
+ function log(message) {
99
+ process.stderr.write(`[workload-picker] ${message}\n`);
100
+ }
101
+
102
+ // ── MCP Server ───────────────────────────────────────────────────────────────
103
+
104
+ const server = new McpServer({
105
+ name: 'workload-picker',
106
+ version: '1.0.0'
107
+ });
108
+
109
+ // Register the list_workloads tool
110
+ server.tool(
111
+ 'list_workloads',
112
+ 'Returns all available benchmark workload profiles with names, descriptions, and use cases',
113
+ {},
114
+ async () => {
115
+ const result = listWorkloads();
116
+ log(`list_workloads → ${result.workloads.length} workloads`);
117
+ return {
118
+ content: [{
119
+ type: 'text',
120
+ text: JSON.stringify(result, null, 2)
121
+ }]
122
+ };
123
+ }
124
+ );
125
+
126
+ // Register the get_workload_profile tool
127
+ server.tool(
128
+ 'get_workload_profile',
129
+ 'Returns benchmark workload parameters for a named workload profile. Use list_workloads first to see available options.',
130
+ {
131
+ workload: z.enum(WORKLOAD_NAMES).describe('Named workload profile to retrieve')
132
+ },
133
+ async ({ workload }) => {
134
+ const profile = getWorkloadProfile(workload);
135
+
136
+ if (!profile) {
137
+ return {
138
+ content: [{
139
+ type: 'text',
140
+ text: JSON.stringify({
141
+ error: `Unknown workload: ${workload}`,
142
+ available: WORKLOAD_NAMES
143
+ })
144
+ }],
145
+ isError: true
146
+ };
147
+ }
148
+
149
+ log(`get_workload_profile → ${workload}`);
150
+ return {
151
+ content: [{
152
+ type: 'text',
153
+ text: JSON.stringify(profile, null, 2)
154
+ }]
155
+ };
156
+ }
157
+ );
158
+
159
+ // ── Start server ─────────────────────────────────────────────────────────────
160
+
161
+ async function main() {
162
+ const transport = new StdioServerTransport();
163
+ log('Starting workload-picker MCP server...');
164
+ await server.connect(transport);
165
+ log('Server connected and ready');
166
+ }
167
+
168
+ main().catch(err => {
169
+ process.stderr.write(`[workload-picker] Fatal startup error: ${err.message}\n`);
170
+ process.exit(1);
171
+ });