Skip to content

KNOWN_ISSUES.md

Latest commit

 

History

History
56 lines (36 loc) · 1.68 KB

KNOWN_ISSUES.md

File metadata and controls

56 lines (36 loc) · 1.68 KB

Known Issues

Performance: Slow API Response with interval='daily' (Issue #609)

Status: Open Priority: P1 Issue: https://github.com/OilpriceAPI/oilpriceapi-api/issues/609

Problem

The /v1/prices/past_year API endpoint with interval='daily' parameter takes 86+ seconds to respond when fetching a full year of data. This can cause timeouts in Kaggle notebooks (120s limit).

Root Cause

The API aggregates 750,000+ intraday tick records into ~363 daily records using Ruby memory aggregation, which is extremely slow.

Impact on Notebooks

  • 01_wti_brent_spread_analysis.ipynb: Fetches WTI + Brent (2 API calls × 86s = 172s) - WILL TIMEOUT
  • 02_oil_price_technical_analysis.ipynb: Fetches Brent only (1 API call × 86s) - MAY TIMEOUT

Workaround Options

Option 1: Reduce Date Range (Recommended)

# Change from 365 days to 180 days (6 months)
start_date = end_date - timedelta(days=180)  # Instead of 365

Option 2: Increase SDK Timeout

# Increase from 120s to 240s
client = OilPriceAPI(api_key=api_key, timeout=240)

Option 3: Run Locally

  • No 120s timeout limit on local machines
  • May take 3-5 minutes total but will complete

Expected Fix

Database-level aggregation using PostgreSQL DATE_TRUNC() will reduce response time from 86s to <2s.

Target: Complete by December 2025

Current Notebook Status

Notebooks are configured for 365 days (1 year) to demonstrate full capabilities, but may timeout on Kaggle.

If you experience timeouts:

  1. Reduce days=365 to days=180 in the date range
  2. Re-run the notebook

Last Updated: 2025-11-26