node-autotask 0.68.0 → 0.69.0
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.
- package/CHANGELOG.md +18 -0
- package/MCP_TOOL_CATALOG.json +2 -2
- package/README.md +51 -10
- package/capabilities.json +7 -7
- package/dist/capabilities-data.js +6 -6
- package/dist/capabilities.cjs +4 -4
- package/dist/capabilities.js +1 -1
- package/dist/{chunk-HFWOHUVF.cjs → chunk-24GRFW55.cjs} +236 -222
- package/dist/{chunk-BF22HBNT.js → chunk-2BB7OASJ.js} +3 -30
- package/dist/{chunk-OUVUQQXH.js → chunk-AH6WAGPS.js} +80 -66
- package/dist/{chunk-CG6OPTFG.cjs → chunk-DPRY3CUJ.cjs} +30 -57
- package/dist/{chunk-V2GE763R.cjs → chunk-JLBSUZ5B.cjs} +993 -395
- package/dist/{chunk-SGH7F2CH.cjs → chunk-KAZE4WC7.cjs} +3 -3
- package/dist/{chunk-DPJLN5BB.js → chunk-QDBJCGQD.js} +986 -390
- package/dist/{chunk-JCMXQUIC.js → chunk-SCFB5MF3.js} +3 -3
- package/dist/{index-DQ08uOH0.d.ts → index-CxBU31jT.d.ts} +16 -9
- package/dist/{index-mIsdpO5d.d.cts → index-WZ3A8nG6.d.cts} +16 -9
- package/dist/index.cjs +92 -85
- package/dist/index.d.cts +3 -3
- package/dist/index.d.ts +3 -3
- package/dist/index.js +14 -7
- package/dist/operations/index.cjs +50 -50
- package/dist/operations/index.d.cts +2 -2
- package/dist/operations/index.d.ts +2 -2
- package/dist/operations/index.js +2 -2
- package/dist/pool.cjs +12 -12
- package/dist/pool.js +4 -4
- package/dist/resources/index.cjs +27 -27
- package/dist/resources/index.d.cts +1 -1
- package/dist/resources/index.d.ts +1 -1
- package/dist/resources/index.js +3 -3
- package/docs/API.md +77 -16
- package/docs/LIVE-TESTING.md +16 -2
- package/docs/VENDOR-QUIRKS.md +54 -0
- package/docs/performance-finance.md +35 -0
- package/examples/mcp-env-config.ts +3 -4
- package/package.json +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -7,6 +7,24 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
|
|
|
7
7
|
|
|
8
8
|
## [Unreleased]
|
|
9
9
|
|
|
10
|
+
## [0.69.0] - 2026-11-17
|
|
11
|
+
|
|
12
|
+
### Breaking
|
|
13
|
+
|
|
14
|
+
- Vendor-protection admission is always enabled: matching SDK clients share a three-request ceiling per object and tracking identifier, and SDK traffic in the same JavaScript realm shares a conservative origin-wide ceiling of 10,000 attempts in a trailing hour (#457). `maxConcurrent` remains the local aggregate bound across objects; `rateLimit.perHour: null` disables only the optional local quota. Independent processes, workers, origins and external integrations require separate coordination.
|
|
15
|
+
|
|
16
|
+
### Changed
|
|
17
|
+
|
|
18
|
+
- Refresh capability test-evidence references for the changed warm-reference and rate-limit scheduling behavior (#457). Operation signatures are unchanged; the evidence update changes the immutable capability-plan hash and requires a minor release.
|
|
19
|
+
- Coalesce equivalent non-fresh record GETs in the SDK read cache, with independent caller cancellation, cloned results and protection against stale in-flight responses after writes or fresh reads (#457). Fresh reads and `cache: false` retain their bypass behavior.
|
|
20
|
+
- Reuse the bounded worker executor for independent composite, name-enrichment and finance branches (#457). Name budgets are reserved before scheduling, results retain input order, failed work drains before returning, and count/preflight guards remain ahead of the walks they authorize.
|
|
21
|
+
|
|
22
|
+
## [0.68.1] - 2026-11-17
|
|
23
|
+
|
|
24
|
+
### Changed
|
|
25
|
+
|
|
26
|
+
- Extract an internal, entity-neutral fuzzy text helper for normalisation policies, fragments, typo scoring, relevance ordering and vocabulary matching (#455). Company live search and its name index share the helper, while company suffixes, field selection, HTTP and cache lifecycle stay in the company adapter. Existing company search contracts and the capability-plan hash are unchanged; future resource adapters can reuse the matching core.
|
|
27
|
+
|
|
10
28
|
## [0.68.0] - 2026-11-17
|
|
11
29
|
|
|
12
30
|
### Added
|
package/MCP_TOOL_CATALOG.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"generatedBy": "scripts/build-tool-catalog.mjs",
|
|
3
|
-
"generatedAt": "2026-11-
|
|
4
|
-
"planHash": "
|
|
3
|
+
"generatedAt": "2026-11-17T12:34:23.940Z",
|
|
4
|
+
"planHash": "959981d335cfd184d8b4e58074d1ab6997b2db87a62683ed13e99c149f8cd4fa",
|
|
5
5
|
"metaTools": [
|
|
6
6
|
{
|
|
7
7
|
"name": "autotask_catalog",
|
package/README.md
CHANGED
|
@@ -175,7 +175,7 @@ server-layer overrides unset rather than blank: an empty value is *invalid* for
|
|
|
175
175
|
| `AUTOTASK_IMPERSONATION_ID` | doctor, examples | optional `ImpersonationResourceId` header |
|
|
176
176
|
| `AUTOTASK_PROFILE` | reference MCP servers | `library` (default) / `agent` — the agent profile's numbers are compared in the template and in *Agent profile* above |
|
|
177
177
|
| `AUTOTASK_RATE_LIMIT_PER_HOUR` | reference MCP servers | hourly limiter: a positive integer, or `0` / `off` to disable |
|
|
178
|
-
| `AUTOTASK_MAX_CONCURRENT` | reference MCP servers | concurrency
|
|
178
|
+
| `AUTOTASK_MAX_CONCURRENT` | reference MCP servers | local aggregate concurrency limit; the SDK separately enforces the shared per-object limit |
|
|
179
179
|
| `AUTOTASK_TIMEOUT_MS` | reference MCP servers | whole-call timeout |
|
|
180
180
|
| `AUTOTASK_MAX_RETRIES` | reference MCP servers | retry budget (`0` disables) |
|
|
181
181
|
| `AUTOTASK_UNTRUSTED_WRAP` | MCP servers | `on` (default) / `off` — the `<untrusted_content>` projection |
|
|
@@ -187,34 +187,75 @@ The SDK itself never reads `process.env`: pass `apiIntegrationCode` / `userName`
|
|
|
187
187
|
explicitly, or call `loadAutotaskCredentialsFromEnv()`.
|
|
188
188
|
|
|
189
189
|
|
|
190
|
+
## Concurrency and vendor budgets
|
|
191
|
+
|
|
192
|
+
`maxConcurrent` is a local aggregate bound per client (library default **10**),
|
|
193
|
+
allowing different objects to progress in parallel. It is not a vendor thread allowance.
|
|
194
|
+
The SDK also enforces a shared ceiling of **3** active requests per object and
|
|
195
|
+
tracking identifier. This gate covers every HTTP verb, query/count and metadata
|
|
196
|
+
reads, and equivalent global/scoped routes. Clients sharing the same normalized
|
|
197
|
+
origin and tracking identifier share the gate across API users, secret rotation
|
|
198
|
+
and impersonation; the tracking identifier is hashed in its internal key.
|
|
199
|
+
ESM and CJS clients share this registry within one JavaScript realm.
|
|
200
|
+
`ZoneInformation` is exempt. There is no setting that raises this ceiling.
|
|
201
|
+
|
|
202
|
+
The vendor applies thread limits to an object and tracking identifier within an
|
|
203
|
+
Autotask instance. New integrations have a limit of 3; some legacy integrations
|
|
204
|
+
may have higher limits. The SDK conservatively uses 3 for both. The vendor also
|
|
205
|
+
adds thread latency starting at 3 simultaneous requests to an object. For
|
|
206
|
+
latency-sensitive work or shared tracking identifiers, lower configured concurrency
|
|
207
|
+
and coordinate traffic with the other callers. The live performance harness uses
|
|
208
|
+
an aggregate cap of 2 to leave room and avoid intentionally entering that latency tier.
|
|
209
|
+
See the official [API thread limits](https://www.autotask.net/help/developerhelp/content/apis/General/ThreadLimiting.htm).
|
|
210
|
+
|
|
211
|
+
The separate **10,000 calls/hour** threshold covers all integrations accessing the
|
|
212
|
+
same database. An always-on SDK guard conservatively caps actual sends to each
|
|
213
|
+
normalized origin at 10,000 over a trailing hour, shared across clients, users,
|
|
214
|
+
tracking identifiers and ESM/CJS loads in the same JavaScript realm. It counts
|
|
215
|
+
retries and failed attempts; cache hits and locally refused work consume no slot.
|
|
216
|
+
The optional `rateLimit.perHour` remains a lower per-client quota;
|
|
217
|
+
`rateLimit: { perHour: null }` disables only that local quota, never the shared
|
|
218
|
+
ceiling. Local `hourlyBudget()` snapshots still describe the per-client quota.
|
|
219
|
+
|
|
220
|
+
An origin can host unrelated databases, so those clients may queue together.
|
|
221
|
+
SDK gates cannot coordinate other processes, worker realms,
|
|
222
|
+
other origins serving the same database, or external integrations, and cannot
|
|
223
|
+
guarantee headroom against that traffic. Check
|
|
224
|
+
`client.thresholdInformation.get()` while headroom remains and coordinate
|
|
225
|
+
deployment-wide budgets. That read bypasses cooldown only; it still obeys and
|
|
226
|
+
consumes hourly admission and the applicable object-thread limit.
|
|
227
|
+
The [vendor threshold reference](https://www.autotask.net/help/developerhelp/content/apis/REST/General_Topics/REST_Thresholds_Limits.htm)
|
|
228
|
+
describes that database-wide limit.
|
|
229
|
+
|
|
190
230
|
## Agent profile (conservative limits for LLM-driven servers)
|
|
191
231
|
|
|
192
232
|
`profile: 'agent'` on the config switches the SDK to conservative limits for
|
|
193
233
|
workloads driven by an LLM: a looping agent can use up the tenant's ~10,000
|
|
194
|
-
req/hour
|
|
195
|
-
concurrent
|
|
196
|
-
|
|
197
|
-
|
|
234
|
+
req/hour database budget (shared across every integration), create excess
|
|
235
|
+
concurrent load, and spend budget on every retry, because the limiter counts
|
|
236
|
+
each attempt. The profile values (plain data): `rateLimit.perHour` **8,000**
|
|
237
|
+
(80% of the vendor allowance; this local setting does not reserve headroom for
|
|
238
|
+
other integrations), page size **25** (capped
|
|
198
239
|
at **100** in agent mode — an explicit value above the cap is clamped, not
|
|
199
240
|
rejected), `timeoutMs` **20,000**, `maxRetries` **1**, `maxConcurrent` **3**,
|
|
200
241
|
`maxResponseBytes` **4 MiB**, resolution scan **200 records / 2 pages**.
|
|
201
242
|
Resolution order: **explicit config > profile > library default** (the
|
|
202
243
|
existing validation runs on the final values); `profile: 'library'` (the
|
|
203
|
-
default)
|
|
204
|
-
"disabled" value that beats the profile's limiter (an absent `perHour` means
|
|
244
|
+
default) keeps its per-client defaults, and `rateLimit: { perHour: null }` is the
|
|
245
|
+
explicit local-quota "disabled" value that beats the profile's limiter (an absent `perHour` means
|
|
205
246
|
"use the profile/default"). Library defaults are unchanged.
|
|
206
247
|
|
|
207
|
-
The hourly
|
|
248
|
+
The optional hourly quota is **per client**: it counts only this client's calls and
|
|
208
249
|
cannot see other instances or integrations on the tenant. The local remaining
|
|
209
250
|
budget is `client.hourlyBudget()` — `{ perHour, used, remaining }`, or
|
|
210
|
-
`undefined` when the
|
|
251
|
+
`undefined` when the local quota is off; the tenant-wide count is
|
|
211
252
|
`client.thresholdInformation.get()`.
|
|
212
253
|
|
|
213
254
|
The reference MCP servers (`examples/mcp-server.ts`,
|
|
214
255
|
`examples/mcp-server-http.ts`) read env overrides through one shared helper
|
|
215
256
|
(`examples/mcp-env-config.ts`; the SDK itself never reads `process.env`):
|
|
216
257
|
`AUTOTASK_PROFILE` (`library` | `agent`), `AUTOTASK_RATE_LIMIT_PER_HOUR`
|
|
217
|
-
(positive integer; `0` / `off` disables the
|
|
258
|
+
(positive integer; `0` / `off` disables the local quota with a stderr warning),
|
|
218
259
|
`AUTOTASK_MAX_CONCURRENT`, `AUTOTASK_TIMEOUT_MS`, `AUTOTASK_MAX_RETRIES`
|
|
219
260
|
(0 disables retries). Unset = the profile's value; an invalid value stops
|
|
220
261
|
startup with an `AutotaskConfigError` naming the variable.
|
package/capabilities.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
|
1
1
|
{
|
|
2
|
-
"generatedAt": "2026-11-
|
|
2
|
+
"generatedAt": "2026-11-17T12:33:45.861Z",
|
|
3
3
|
"groups": {
|
|
4
4
|
"B": {
|
|
5
|
-
"planHash": "
|
|
6
|
-
"builtAt": "2026-11-
|
|
5
|
+
"planHash": "959981d335cfd184d8b4e58074d1ab6997b2db87a62683ed13e99c149f8cd4fa",
|
|
6
|
+
"builtAt": "2026-11-17T12:33:45.837Z",
|
|
7
7
|
"records": [
|
|
8
8
|
{
|
|
9
9
|
"id": "appointments.count",
|
|
@@ -42786,8 +42786,8 @@
|
|
|
42786
42786
|
]
|
|
42787
42787
|
},
|
|
42788
42788
|
"A": {
|
|
42789
|
-
"planHash": "
|
|
42790
|
-
"builtAt": "2026-11-
|
|
42789
|
+
"planHash": "959981d335cfd184d8b4e58074d1ab6997b2db87a62683ed13e99c149f8cd4fa",
|
|
42790
|
+
"builtAt": "2026-11-17T12:33:45.837Z",
|
|
42791
42791
|
"records": [
|
|
42792
42792
|
{
|
|
42793
42793
|
"id": "companies.count",
|
|
@@ -82294,8 +82294,8 @@
|
|
|
82294
82294
|
]
|
|
82295
82295
|
},
|
|
82296
82296
|
"operations": {
|
|
82297
|
-
"planHash": "
|
|
82298
|
-
"builtAt": "2026-11-
|
|
82297
|
+
"planHash": "959981d335cfd184d8b4e58074d1ab6997b2db87a62683ed13e99c149f8cd4fa",
|
|
82298
|
+
"builtAt": "2026-11-17T12:33:45.837Z",
|
|
82299
82299
|
"records": [
|
|
82300
82300
|
{
|
|
82301
82301
|
"id": "operations.aggregate",
|
|
@@ -348,8 +348,8 @@ const CAPABILITY_DATA = {
|
|
|
348
348
|
],
|
|
349
349
|
"groups": {
|
|
350
350
|
"B": {
|
|
351
|
-
"planHash": "
|
|
352
|
-
"builtAt": "2026-11-
|
|
351
|
+
"planHash": "959981d335cfd184d8b4e58074d1ab6997b2db87a62683ed13e99c149f8cd4fa",
|
|
352
|
+
"builtAt": "2026-11-17T12:33:45.837Z",
|
|
353
353
|
"records": [
|
|
354
354
|
{
|
|
355
355
|
"id": "appointments.count",
|
|
@@ -43132,8 +43132,8 @@ const CAPABILITY_DATA = {
|
|
|
43132
43132
|
]
|
|
43133
43133
|
},
|
|
43134
43134
|
"A": {
|
|
43135
|
-
"planHash": "
|
|
43136
|
-
"builtAt": "2026-11-
|
|
43135
|
+
"planHash": "959981d335cfd184d8b4e58074d1ab6997b2db87a62683ed13e99c149f8cd4fa",
|
|
43136
|
+
"builtAt": "2026-11-17T12:33:45.837Z",
|
|
43137
43137
|
"records": [
|
|
43138
43138
|
{
|
|
43139
43139
|
"id": "companies.count",
|
|
@@ -82640,8 +82640,8 @@ const CAPABILITY_DATA = {
|
|
|
82640
82640
|
]
|
|
82641
82641
|
},
|
|
82642
82642
|
"operations": {
|
|
82643
|
-
"planHash": "
|
|
82644
|
-
"builtAt": "2026-11-
|
|
82643
|
+
"planHash": "959981d335cfd184d8b4e58074d1ab6997b2db87a62683ed13e99c149f8cd4fa",
|
|
82644
|
+
"builtAt": "2026-11-17T12:33:45.837Z",
|
|
82645
82645
|
"records": [
|
|
82646
82646
|
{
|
|
82647
82647
|
"id": "operations.aggregate",
|
package/dist/capabilities.cjs
CHANGED
|
@@ -1,18 +1,18 @@
|
|
|
1
1
|
'use strict';
|
|
2
2
|
|
|
3
|
-
var
|
|
3
|
+
var chunkKAZE4WC7_cjs = require('./chunk-KAZE4WC7.cjs');
|
|
4
4
|
|
|
5
5
|
|
|
6
6
|
|
|
7
7
|
Object.defineProperty(exports, "CAPABILITY_PLAN_HASHES", {
|
|
8
8
|
enumerable: true,
|
|
9
|
-
get: function () { return
|
|
9
|
+
get: function () { return chunkKAZE4WC7_cjs.CAPABILITY_PLAN_HASHES; }
|
|
10
10
|
});
|
|
11
11
|
Object.defineProperty(exports, "loadCapabilities", {
|
|
12
12
|
enumerable: true,
|
|
13
|
-
get: function () { return
|
|
13
|
+
get: function () { return chunkKAZE4WC7_cjs.loadCapabilities; }
|
|
14
14
|
});
|
|
15
15
|
Object.defineProperty(exports, "loadCapabilityRecords", {
|
|
16
16
|
enumerable: true,
|
|
17
|
-
get: function () { return
|
|
17
|
+
get: function () { return chunkKAZE4WC7_cjs.loadCapabilityRecords; }
|
|
18
18
|
});
|
package/dist/capabilities.js
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
export { CAPABILITY_PLAN_HASHES, loadCapabilities, loadCapabilityRecords } from './chunk-
|
|
1
|
+
export { CAPABILITY_PLAN_HASHES, loadCapabilities, loadCapabilityRecords } from './chunk-SCFB5MF3.js';
|