feat: Add error ratio-based circuit breaking policy to api-breaker plugin #12765
Conversation
…ugin - Add new 'unhealthy-ratio' policy that triggers circuit breaker based on error rate within sliding time window - Implement three-state circuit breaker: CLOSED -> OPEN -> HALF_OPEN -> CLOSED - Add configurable parameters: error_ratio, min_request_threshold, sliding_window_size, permitted_number_of_calls_in_half_open_state, success_ratio - Maintain full backward compatibility with existing 'unhealthy-count' policy as default - Add comprehensive test coverage for new functionality - Update documentation in both Chinese and English - Follow APISIX coding standards and testing conventions This enhancement provides more intelligent circuit breaking for microservices architectures by considering error rates rather than just consecutive failure counts.
dosubot
bot
added
size:XXL
Baoyuantop
left a comment
Baoyuantop
left a comment
There was a problem hiding this comment.
Thanks for your contribution! Based on the current configuration, we need to add some test cases:
-
After the sliding window time (sliding_window_size) expires, are the statistics (total number of requests, number of failures) correctly cleared?
-
Failure fallback in half-open state (Half-Open -> Open)
-
Sending more requests than permitted_number_of_calls_in_half_open_state in half-open state
| description = "Size of the sliding window in seconds" | ||
| }, | ||
| default = {http_statuses = {500}, failures = 3} | ||
| permitted_number_of_calls_in_half_open_state = { |
There was a problem hiding this comment.
Is it possible to optimize this variable name?
| === TEST $((${1}+1)): hit route (return 200) | ||
| --- request | ||
| GET /api_breaker | ||
| --- response_body | ||
| hello world | ||
|
|
||
|
|
||
|
|
||
| === TEST $((${1}+1)): hit route and return 500 (first failure) | ||
| --- request | ||
| GET /api_breaker?code=500 | ||
| --- error_code: 500 | ||
| --- response_body | ||
| fault injection! |
There was a problem hiding this comment.
We can make multiple requests in a single case; you can refer to the tests in api-breaker.t
There was a problem hiding this comment.
May I ask if there is an official test image of apisix? It is very difficult to set up the environment for testing .t files locally
There was a problem hiding this comment.
2 participants
feat: Add error ratio-based circuit breaking policy to api-breaker plugin
What this PR does / why we need it
This PR implements error ratio-based circuit breaking (
unhealthy-ratiopolicy) for theapi-breakerplugin, providing more intelligent and adaptive circuit breaking behavior based on error rates within a sliding time window, rather than just consecutive failure counts.Closes #12763
Types of changes
Description
Current Limitations
New Features Added
unhealthy-ratiopolicy that triggers circuit breaker based on error rate within a sliding time windowNew Configuration Parameters
policy"unhealthy-count"unhealthy.error_ratio0.5unhealthy.min_request_threshold10unhealthy.sliding_window_size300unhealthy.permitted_number_of_calls_in_half_open_state3healthy.success_ratio0.6Example Configuration
{ "plugins": { "api-breaker": { "break_response_code": 503, "policy": "unhealthy-ratio", "max_breaker_sec": 60, "unhealthy": { "http_statuses": [500, 502, 503, 504], "error_ratio": 0.5, "min_request_threshold": 10, "sliding_window_size": 300, "permitted_number_of_calls_in_half_open_state": 3 }, "healthy": { "http_statuses": [200, 201, 202], "success_ratio": 0.6 } } } }How Has This Been Tested?
Test Results
Files Modified
apisix/plugins/api-breaker.lua- Core plugin logic with new ratio-based policyt/plugin/api-breaker2.t- New comprehensive test file for ratio-based circuit breakingdocs/en/latest/plugins/api-breaker.md- Updated English documentationdocs/zh/latest/plugins/api-breaker.md- Updated Chinese documentationChecklist
Additional Notes
This implementation:
The feature addresses real-world use cases for:
Ready for review and feedback!