🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.
TypeScript analyticsautogenevaluationlangchainlarge-language-modelsllama-indexllmllm-evaluationllm-observabilityllmopsmonitoringobservabilityopen-sourceopenaiplaygroundprompt-engineeringprompt-managementself-hostedycombinator
import { TRACE_DOWNLOAD_OMIT_LARGE_FIELDS_THRESHOLD } from "@/src/features/traces/constants/traceDownloadConfig";
/**
* Configuration constants for trace views.
*
* Centralizes thresholds, dimensions, and behavior settings
* for easier tuning and maintenance.
*/
export const TRACE_VIEW_CONFIG = {
/**
* Log View configuration
*/
logView: {
/**
* Number of observations above which virtualization is enabled.
* Virtualization renders only visible rows for better performance.
*/
virtualizationThreshold: 350,
/**
* Number of observations above which download uses cached I/O only.
* Above this threshold, only expanded observations include full I/O data.
*/
downloadThreshold: TRACE_DOWNLOAD_OMIT_LARGE_FIELDS_THRESHOLD,
/**
* Row heights for virtualization calculations
*/
rowHeight: {
collapsed: 28,
expanded: 150,
},
/**
* Maximum tree depth for indent visualization.
* Deeper trees disable indent to prevent excessive horizontal space.
*/
maxIndentDepth: 5,
/**
* Indent size per level (in pixels)
*/
indentPx: 12,
/**
* Prefetching behavior for viewport-based observation loading
*/
prefetch: {
/**
* Margin around viewport for triggering prefetch (e.g., "100px")
* Observations within this margin will be prefetched.
*/
rootMargin: "100px",
/**
* Debounce delay in milliseconds before triggering prefetch.
* Prevents request storms during fast scrolling.
*/
debounceMs: 250,
},
/**
* Batch fetching configuration for loading all observations
*/
batchFetch: {
/**
* Maximum number of concurrent observation fetch requests
*/
concurrency: 10,
},
},
} as const;