Skip to content

BITVAVO_RATE_LIMIT_FIX.md

Latest commit

 

History

History
154 lines (111 loc) · 5.09 KB

BITVAVO_RATE_LIMIT_FIX.md

File metadata and controls

154 lines (111 loc) · 5.09 KB
Raw

Bitvavo Rate Limit Fix

Problem

The python-bitvavo-api library (version 1.4.3 and earlier) has a bug in its rate limiting thread that causes the following error:

Exception in thread Thread-1:
Traceback (most recent call last):
  File "/usr/lib/python3.12/threading.py", line 1073, in _bootstrap_inner
    self.run()
  File ".../python_bitvavo_api/bitvavo.py", line 112, in run
    self.waitForReset(self.timeToWait)
  File ".../python_bitvavo_api/bitvavo.py", line 102, in waitForReset
    time.sleep(waitTime)
ValueError: sleep length must be non-negative

Root Cause

The bug occurs in the rateLimitThread.waitForReset() method when calculating the wait time:

timeToWait = (self.bitvavo.rateLimitReset / 1000) - time.time()

If the current time is already past the rateLimitReset timestamp (meaning the rate limit ban has already expired), this calculation produces a negative value, which causes time.sleep() to raise a ValueError.

This typically happens when:

  • System time jumps forward (clock sync, timezone change)
  • Network delays cause the rate limit check to occur after the ban expired
  • High system load causes thread scheduling delays

Solution

We've implemented a monkey patch in src/trader/bitvavo.py that fixes this issue by:

  1. Checking for negative wait times before calling time.sleep()
  2. Converting negative values to zero (no wait needed if ban already expired)
  3. Adding comprehensive logging to monitor rate limit events
  4. Handling recursive calls to ensure all paths are protected

Implementation

The patch is automatically applied when the BitvavoClient is imported:

def _patch_bitvavo_rate_limit():
    """Monkey patch the Bitvavo library to fix the negative sleep bug."""
    from python_bitvavo_api.bitvavo import rateLimitThread

    def patched_wait_for_reset(self, waitTime):
        # Ensure non-negative wait time
        if waitTime < 0:
            logger.warning(f"Negative wait time detected ({waitTime:.2f}s). Setting to 0.")
            waitTime = 0

        time.sleep(waitTime)

        # Handle recursive calls with negative time protection
        if time.time() >= self.bitvavo.rateLimitReset:
            timeToWait = max(0, (self.bitvavo.rateLimitReset / 1000) - time.time())
            if timeToWait > 0:
                self.waitForReset(timeToWait)

    rateLimitThread.waitForReset = patched_wait_for_reset

# Apply patch on import
_patch_bitvavo_rate_limit()

Logging

The patch adds detailed logging at the trader.bitvavo logger level:

  • INFO: Rate limit waits and resets
  • WARNING: Negative wait times detected
  • ERROR: Patch application failures

Example log output:

INFO - BitvavoClient initialized with rate limit protection
WARNING - Bitvavo rate limit: Negative wait time detected (-5.23s). Ban likely already expired. Setting to 0.
INFO - Bitvavo rate limit: Waiting 45.2s for rate limit reset
INFO - Bitvavo rate limit: Ban lifted, reset to 1000 requests

Testing

Comprehensive unit tests are available in tests/test_bitvavo_patch.py:

pytest tests/test_bitvavo_patch.py -v

Test coverage includes:

  • ✅ Negative wait time handling
  • ✅ Positive wait time (unchanged behavior)
  • ✅ Zero wait time
  • ✅ Recursive calls with negative times
  • ✅ Client initialization with patch
  • ✅ Patch application verification

Status

✅ FIXED - The issue is resolved in Bender's codebase

Version Information

  • Library: python-bitvavo-api version 1.4.3 (latest as of 2025-10-23)
  • Bug Status: Still present in upstream library
  • Our Fix: Monkey patch applied automatically on import
  • Alternative: Wait for upstream fix (may not happen)

Upstream Issue

This is a bug in the external python-bitvavo-api library, not in Bender's code. Consider:

  1. Reporting to maintainers: Create an issue at https://github.com/bitvavo/python-bitvavo-api
  2. Pull request: Submit a fix to the upstream repository
  3. Monitor for updates: Check if future versions fix this issue

Migration Path

If the upstream library fixes this bug in a future version:

  1. Update python-bitvavo-api to the fixed version
  2. Test with pytest tests/test_bitvavo_patch.py
  3. The patch will remain safe even if the library is fixed (it just becomes redundant)
  4. Optionally remove the patch if no longer needed

Impact

Before Fix:

  • ❌ Random crashes in rate limit thread
  • ❌ No visibility into rate limit events
  • ❌ Potential trading interruptions

After Fix:

  • ✅ No more ValueError: sleep length must be non-negative
  • ✅ Comprehensive logging of rate limit events
  • ✅ Graceful handling of expired bans
  • ✅ Stable 24/7 operation

Additional Notes

  • The patch is non-invasive and doesn't affect normal operation
  • Performance impact is negligible (just an if check and max() call)
  • The fix is backward compatible with older python-bitvavo-api versions
  • No configuration changes needed - works automatically

Last Updated: 2025-10-23 Tested With: python-bitvavo-api 1.4.3