feat(cli): add ingest LLM rate-limit governor with paced retries (#261)

* feat(cli): add ingest rate limit governor

* feat(cli): wire ingest rate-limit config

* feat(cli): report provider rate-limit signals

* feat(cli): show ingest rate-limit waits

* fix(cli): complete rate-limit event coverage

* fix(cli): abort ingest provider calls cleanly

* fix(cli): propagate ingest cancellation

* fix(cli): reject pre-aborted ingest rate-limit waits

* fix(cli): honor Claude rate-limit reset waits

* fix(cli): retry thrown Codex rate-limit failures

* fix(cli): type Claude rate-limit result details

* fix(cli): emit ingest rate-limit countdowns from rejected signals

* fix(cli): report ai sdk rate-limit header utilization

* fix(cli): gate LLM rate-limit retries on the governor budget

The AI SDK and Codex runtimes retried 429 / opaque rate-limit failures up
to 6-7 times with no backoff when constructed without a RateLimitGovernor
(scan, memory, setup) or with pacing disabled, ignoring Retry-After and
worsening the limit. The outer retry loop only cooperates with the
governor's pause, so without active pacing there is no backoff to apply.

Route the retry bound through a single source: RateLimitGovernor
.maxRetryAttempts(), which returns retry.maxAttempts when enabled and 1
(no outer retry) when absent or disabled. All three runtimes (ai-sdk,
codex, claude-code) now use it, so ingest.rateLimit.retry.maxAttempts
genuinely controls attempts and the hard-coded 6 (plus Codex's off-by-one
extra attempt) is gone. Backend-native retry (e.g. the AI SDK's maxRetries)
still handles transient 429s.

Also correct the ktx.yaml docs for maxWaitMs (caps each wait, not the whole
run) and maxAttempts, and sync uv.lock ktx-sl/ktx-daemon to 0.9.0.
This commit is contained in:
Andrey Avtomonov 2026-06-05 12:10:27 +02:00 committed by GitHub
parent 5a8821073b
commit c3d8cedb0b
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
35 changed files with 2336 additions and 72 deletions

View file

@ -12,6 +12,7 @@ import type { KtxSemanticLayerComputePort } from '../../context/daemon/semantic-
import { createRuntimeToolDescriptorFromAiTool } from '../../context/llm/runtime-tools.js';
import { createLocalKtxLlmRuntimeFromConfig } from '../../context/llm/local-config.js';
import { KtxIngestEmbeddingPortAdapter } from '../../context/llm/embedding-port.js';
import { createRateLimitGovernorConfig, RateLimitGovernor } from '../../context/llm/rate-limit-governor.js';
import { RuntimeAgentRunner, type AgentRunnerPort, type KtxLlmRuntimePort, type KtxRuntimeToolSet } from '../../context/llm/runtime-port.js';
import type { KtxEmbeddingProvider } from '../../llm/types.js';
import type { KtxLocalProject } from '../../context/project/project.js';
@ -619,7 +620,7 @@ function localIngestLlmProviderGuardMessage(projectDir: string): string {
].join('\n');
}
function resolveAgentRunner(options: CreateLocalBundleIngestRuntimeOptions): {
function resolveAgentRunner(options: CreateLocalBundleIngestRuntimeOptions, rateLimitGovernor: RateLimitGovernor): {
agentRunner: AgentRunnerPort;
llmRuntime?: KtxLlmRuntimePort;
} {
@ -628,6 +629,7 @@ function resolveAgentRunner(options: CreateLocalBundleIngestRuntimeOptions): {
(options.createLlmRuntime ?? createLocalKtxLlmRuntimeFromConfig)(options.project.config.llm, {
projectDir: options.project.projectDir,
env: process.env,
rateLimitGovernor,
}) ??
undefined;
@ -677,7 +679,13 @@ export function createLocalBundleIngestRuntime(
const knowledgeIndex = new LocalKnowledgeIndex(options.project, embedding);
const knowledgeEvents = new NoopKnowledgeEventPort();
const wikiService = new KnowledgeWikiService(rootFileStore, embedding, knowledgeIndex, options.project.git, logger);
const { agentRunner, llmRuntime } = resolveAgentRunner(options);
const rateLimitGovernor = new RateLimitGovernor(
createRateLimitGovernorConfig({
...options.project.config.ingest.rateLimit,
maxConcurrency: options.project.config.ingest.workUnits.maxConcurrency,
}),
);
const { agentRunner, llmRuntime } = resolveAgentRunner(options, rateLimitGovernor);
const promptService = new PromptService({ promptsDir, partials: [], logger });
const storage = new LocalIngestStorage(options.project);
const registry = registerAdapters(options.adapters);
@ -717,6 +725,7 @@ export function createLocalBundleIngestRuntime(
workUnitMaxConcurrency: options.project.config.ingest.workUnits.maxConcurrency,
workUnitStepBudget: options.project.config.ingest.workUnits.stepBudget,
workUnitFailureMode: options.project.config.ingest.workUnits.failureMode,
rateLimitGovernor,
profileIngest: options.project.config.ingest.profile,
ingestTraceLevel: ingestTraceLevelFromEnv(),
},