--- title: "Configuration" description: "Control log level, connectors, file rotation, ingestion paths, and config loading with code or blyp.config files." canonical_url: "https://www.blyp.dev/docs/configuration" markdown_url: "https://www.blyp.dev/docs/configuration.md" last_updated: "2018-10-20" x_farming_labs_generated_preamble: true --- # Configuration URL: /docs/configuration LLM index: /llms.txt Description: Control log level, connectors, file rotation, ingestion paths, and config loading with code or blyp.config files. # Configuration Blyp configuration comes from two places: - `blyp.config.*` in your project root - per-logger runtime overrides passed to `createStandaloneLogger()` or a framework adapter Runtime overrides are merged on top of the loaded config. ## Supported config files Blyp loads the first matching config file from: - `blyp.config.ts` - `blyp.config.mts` - `blyp.config.cts` - `blyp.config.js` - `blyp.config.mjs` - `blyp.config.cjs` - `blyp.config.json` Executable config files are loaded through `jiti`. If Convex (or another isolate) will import the same `blyp.config.ts`, use the isolate-safe helper so that file does not pull the Node logger graph: ```ts import { defineConfig } from "@blyp/core/config"; export default defineConfig({ level: "info", connectors: { otlp: [ { name: "axiom", endpoint: "https://api.axiom.co/v1/logs", auth: `Bearer ${process.env.AXIOM_TOKEN}`, }, ], }, }); ``` Node apps can keep `import { defineConfig } from "@blyp/core"`. Convex must import the config module and pass it to `configureConvexLogger()`; it does not walk the filesystem. See [Convex](/docs/integrations/convex). ## Example `blyp.config.ts` ```ts export default { "pretty": true, "level": "info", "destination": "database", "redact": { "paths": ["user.ssn", "payment.**.raw"] }, "file": { "enabled": true, "format": "ndjson", "rotation": { "enabled": true, "maxSizeBytes": 10485760, "maxArchives": 5, "compress": true } }, "database": { "dialect": "postgres", "delivery": { "strategy": "immediate", "batchSize": 1, "flushIntervalMs": 250, "maxQueueSize": 1000, "overflowStrategy": "drop-oldest", "flushTimeoutMs": 5000, "retry": { "maxRetries": 1, "backoffMs": 100 } } }, "clientLogging": { "enabled": true, "path": "/inngest" }, "connectors": { "delivery": { "enabled": true, "durableQueuePath": ".blyp/connectors.sqlite", "retry": { "maxAttempts": 8, "initialBackoffMs": 500, "maxBackoffMs": 30000, "multiplier": 2, "jitter": true } }, "posthog": { "enabled": true, "mode": "auto", "projectKey": process.env.POSTHOG_PROJECT_KEY, "errorTracking": { "enabled": true, "mode": "auto" } }, "sentry": { "enabled": true, "mode": "auto", "dsn": process.env.SENTRY_DSN }, "otlp": [ { "name": "grafana", "enabled": true, "mode": "auto", "endpoint": process.env.OTLP_HTTP_ENDPOINT } ] } }; ``` ## Default behavior - pretty console logging is enabled - default level is `info` - default destination is `file` - file logging is enabled - client logging is enabled - default client ingestion path is `/inngest` - built-in redaction runs across console, files, database writes, connectors, client ingestion, and framework request logs - default rotation is `10 MB`, `5` archives, gzip compression enabled - database delivery defaults are immediate writes, queue size `1000`, flush timeout `5000ms`, and one retry with `100ms` backoff - connector delivery queueing is disabled unless `connectors.delivery.enabled` is set - connectors are disabled unless configured ## Redaction config Use `redact` to remove sensitive values before they reach Blyp sinks: ```ts export default { redact: { keys: ["internal_token"], paths: ["user.ssn", "payment.**.raw"], patterns: [/MY_ORG_[A-Z0-9]{32}/], disablePatternScanning: false, }, }; ``` The redaction config supports: - `keys`: additional sensitive field names to redact - `paths`: exact or wildcard field paths to redact - `patterns`: regex patterns for secret scanning - `disablePatternScanning`: disables Blyp's built-in token and card pattern detection Notes: - `redact.paths` supports exact paths, numeric indexes, `*`, and `**` - `patterns` require executable config such as `blyp.config.ts` - Blyp preserves keys and replaces values with `[REDACTED]` or a typed marker such as `[REDACTED:jwt]` ## Database logging config Use `destination: "database"` when local file persistence is not a good fit, especially in serverless deployments. Database mode requires an executable config file such as `blyp.config.ts`, `blyp.config.mts`, `blyp.config.js`, or `blyp.config.cjs` because Prisma and Drizzle adapters are runtime objects. `blyp.config.json` cannot hold adapter instances by itself. The config flag is only one part of the setup. The database must also match the required Blyp schema contract documented in [Schema Contract](/docs/database/schema). ```ts import { PrismaClient } from "@prisma/client"; import { createPrismaDatabaseAdapter } from "@blyp/core/database"; const prisma = new PrismaClient(); export default { destination: "database", database: { dialect: "postgres", adapter: createPrismaDatabaseAdapter({ client: prisma, model: "blypLog", }), delivery: { strategy: "batch", batchSize: 50, flushIntervalMs: 1000, }, }, }; ``` The database delivery config supports: - `strategy`: `"immediate"` or `"batch"` - `batchSize` - `flushIntervalMs` - `maxQueueSize` - `overflowStrategy`: `"drop-oldest"` or `"drop-new"` - `flushTimeoutMs` - `retry.maxRetries` - `retry.backoffMs` Use [Database](/docs/database) for the full setup flow, [Schema Contract](/docs/database/schema) for the required table and indexes, and [Database Migrations](/docs/database/migrations) for the CLI migration workflow. ## Connector delivery queue config Use `connectors.delivery` when retryable server-side connector failures should spill into Blyp's internal queue instead of being dropped immediately: ```ts export default { connectors: { delivery: { enabled: true, durableQueuePath: ".blyp/connectors.sqlite", memoryBufferSize: 500, durableSpillStrategy: "after-first-failure", memoryBatchSize: 25, sqliteWriteBatchSize: 100, sqliteReadBatchSize: 50, dispatchConcurrency: 4, pollIntervalMs: 1000, overflowStrategy: "drop-oldest", retry: { maxAttempts: 8, initialBackoffMs: 500, maxBackoffMs: 30000, multiplier: 2, jitter: true, }, }, betterstack: { enabled: true, mode: "auto", sourceToken: process.env.BETTERSTACK_TOKEN, ingestingHost: "https://in.logs.betterstack.com", }, }, }; ``` The connector delivery config supports: - `enabled` - `memoryBufferSize` - `durableQueuePath` - `durableSpillStrategy` - `memoryBatchSize` - `sqliteWriteBatchSize` - `sqliteReadBatchSize` - `dispatchConcurrency` - `pollIntervalMs` - `overflowStrategy` - `retry.maxAttempts` - `retry.initialBackoffMs` - `retry.maxBackoffMs` - `retry.multiplier` - `retry.jitter` Notes: - the default queue path is `.blyp/connectors.sqlite` - `.blyp/` should be ignored in git - this queue applies to server connector delivery, not the primary `destination` - older runtimes without built-in SQLite support fall back to memory-only retries with a warning - use the root exports `getConnectorQueuePath()`, `connectorQueueExists()`, `getConnectorDeliveryStatusSummary()`, `listConnectorDeadLetters()`, `retryConnectorDeadLetters()`, and `clearConnectorDeadLetters()` for queue inspection and recovery tooling ## Common server adapter options ```ts createLogger({ level: "debug", pretty: false, logDir: "logs/api", file: { rotation: { maxSizeBytes: 5 * 1024 * 1024, maxArchives: 3, compress: true, }, }, autoLogging: { ignore: (ctx) => false, }, customProps: (ctx) => ({ requestId: "req_123", }), logErrors: true, ignorePaths: ["/health", "/metrics/**"], clientLogging: { path: "/client-ingest", validate: async (_ctx, payload) => payload.id !== "blocked", enrich: async (_ctx, payload) => ({ clientLevel: payload.level, }), }, }); ``` ## Connector config Connectors live under `connectors`: ```ts export default { connectors: { posthog: { enabled: true, mode: "auto", projectKey: process.env.POSTHOG_PROJECT_KEY, errorTracking: { enabled: true, mode: "auto", enableExceptionAutocapture: true, }, }, sentry: { enabled: true, mode: "auto", dsn: process.env.SENTRY_DSN, environment: process.env.NODE_ENV, release: process.env.RELEASE_SHA, }, otlp: [ { name: "grafana", enabled: true, mode: "auto", endpoint: process.env.OTLP_HTTP_ENDPOINT, auth: process.env.OTLP_AUTH_HEADER, serviceName: "api", }, ], }, }; ``` ## Notes that matter in production - `ignorePaths` applies to both automatic request logs and error logs. - When client logging is enabled, the ingestion path is auto-added to the ignore list so Blyp does not emit duplicate `http_request` records for `POST /inngest`. - `logErrors: false` disables HTTP error record emission for the adapter. - `customProps()` is merged into emitted request records and is the right place for request IDs, user IDs, tenant IDs, and controller/action names. - connector `auto` mode forwards from server runtimes directly - browser and Expo connector requests still flow through Blyp ingestion first ## Config helpers from the root package ```ts import { getConfig, loadConfig, mergeBlypConfig, resolveConfig, resetConfigCache, type ResolvedBlypConfig, } from "@blyp/core"; ``` `loadConfig()`, `resolveConfig()`, and `getConfig()` now return `ResolvedBlypConfig`, so the returned shape includes normalized defaults for `destination`, `file`, `clientLogging`, connector readiness, and database delivery. Use these if you need to inspect or recompute configuration in tests or custom bootstrap code. Use [Connectors](/docs/connectors) for provider-specific behavior and manual APIs. ## Sitemap See the full [sitemap](/sitemap.md) for all pages. Docs-scoped sitemap: [/docs/sitemap.md](/docs/sitemap.md). Well-known sitemap: [/.well-known/sitemap.md](/.well-known/sitemap.md).