support set async circuit breaker for memcached and redis client (#5789)

* support set async circuit breaker for memcached and redis client

Signed-off-by: Ben Ye <[email protected]>

* changelog

Signed-off-by: Ben Ye <[email protected]>

* docs

Signed-off-by: Ben Ye <[email protected]>

---------

Signed-off-by: Ben Ye <[email protected]>

Author: yeya24

CHANGELOG.md

@@ -13,6 +13,7 @@
 * [FEATURE] Tracing: Add `tracing.otel.round-robin` flag to use `round_robin` gRPC client side LB policy for sending OTLP traces. #5731
 * [FEATURE] Ruler: Add `ruler.concurrent-evals-enabled` flag to enable concurrent evaluation within a single rule group for independent rules. Maximum concurrency can be configured via `ruler.max-concurrent-evals`. #5766
 * [FEATURE] Distributor Queryable: Experimental: Add config `zone_results_quorum_metadata`. When querying ingesters using metadata APIs such as label names and values, only results from quorum number of zones will be included and merged. #5779
+* [FEATURE] Storage Cache Clients: Add config `set_async_circuit_breaker_config` to utilize the circuit breaker pattern for dynamically thresholding asynchronous set operations. Implemented in both memcached and redis cache clients. #5789
 * [ENHANCEMENT] Store Gateway: Added `-store-gateway.enabled-tenants` and `-store-gateway.disabled-tenants` to explicitly enable or disable store-gateway for specific tenants. #5638
 * [ENHANCEMENT] Compactor: Add new compactor metric `cortex_compactor_start_duration_seconds`. #5683
 * [ENHANCEMENT] Upgraded Docker base images to `alpine:3.18`. #5684

docs/blocks-storage/querier.md

@@ -606,6 +606,35 @@ blocks_storage:
         # CLI flag: -blocks-storage.bucket-store.index-cache.memcached.auto-discovery
         [auto_discovery: <boolean> | default = false]
 
+        set_async_circuit_breaker_config:
+          # If true, enable circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.memcached.set-async.circuit-breaker.enabled
+          [enabled: <boolean> | default = false]
+
+          # Maximum number of requests allowed to pass through when the circuit
+          # breaker is half-open. If set to 0, by default it allows 1 request.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.memcached.set-async.circuit-breaker.half-open-max-requests
+          [half_open_max_requests: <int> | default = 10]
+
+          # Period of the open state after which the state of the circuit
+          # breaker becomes half-open. If set to 0, by default open duration is
+          # 60 seconds.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.memcached.set-async.circuit-breaker.open-duration
+          [open_duration: <duration> | default = 5s]
+
+          # Minimal requests to trigger the circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.memcached.set-async.circuit-breaker.min-requests
+          [min_requests: <int> | default = 50]
+
+          # Consecutive failures to determine if the circuit breaker should
+          # open.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.memcached.set-async.circuit-breaker.consecutive-failures
+          [consecutive_failures: <int> | default = 5]
+
+          # Failure percentage to determine if the circuit breaker should open.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.memcached.set-async.circuit-breaker.failure-percent
+          [failure_percent: <float> | default = 0.05]
+
         # Selectively cache index item types. Supported values are Postings,
         # ExpandedPostings and Series
         # CLI flag: -blocks-storage.bucket-store.index-cache.memcached.enabled-items
@@ -707,6 +736,35 @@ blocks_storage:
         # CLI flag: -blocks-storage.bucket-store.index-cache.redis.cache-size
         [cache_size: <int> | default = 0]
 
+        set_async_circuit_breaker_config:
+          # If true, enable circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.redis.set-async.circuit-breaker.enabled
+          [enabled: <boolean> | default = false]
+
+          # Maximum number of requests allowed to pass through when the circuit
+          # breaker is half-open. If set to 0, by default it allows 1 request.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.redis.set-async.circuit-breaker.half-open-max-requests
+          [half_open_max_requests: <int> | default = 10]
+
+          # Period of the open state after which the state of the circuit
+          # breaker becomes half-open. If set to 0, by default open duration is
+          # 60 seconds.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.redis.set-async.circuit-breaker.open-duration
+          [open_duration: <duration> | default = 5s]
+
+          # Minimal requests to trigger the circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.redis.set-async.circuit-breaker.min-requests
+          [min_requests: <int> | default = 50]
+
+          # Consecutive failures to determine if the circuit breaker should
+          # open.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.redis.set-async.circuit-breaker.consecutive-failures
+          [consecutive_failures: <int> | default = 5]
+
+          # Failure percentage to determine if the circuit breaker should open.
+          # CLI flag: -blocks-storage.bucket-store.index-cache.redis.set-async.circuit-breaker.failure-percent
+          [failure_percent: <float> | default = 0.05]
+
         # Selectively cache index item types. Supported values are Postings,
         # ExpandedPostings and Series
         # CLI flag: -blocks-storage.bucket-store.index-cache.redis.enabled-items
@@ -779,6 +837,35 @@ blocks_storage:
         # CLI flag: -blocks-storage.bucket-store.chunks-cache.memcached.auto-discovery
         [auto_discovery: <boolean> | default = false]
 
+        set_async_circuit_breaker_config:
+          # If true, enable circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.memcached.set-async.circuit-breaker.enabled
+          [enabled: <boolean> | default = false]
+
+          # Maximum number of requests allowed to pass through when the circuit
+          # breaker is half-open. If set to 0, by default it allows 1 request.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.memcached.set-async.circuit-breaker.half-open-max-requests
+          [half_open_max_requests: <int> | default = 10]
+
+          # Period of the open state after which the state of the circuit
+          # breaker becomes half-open. If set to 0, by default open duration is
+          # 60 seconds.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.memcached.set-async.circuit-breaker.open-duration
+          [open_duration: <duration> | default = 5s]
+
+          # Minimal requests to trigger the circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.memcached.set-async.circuit-breaker.min-requests
+          [min_requests: <int> | default = 50]
+
+          # Consecutive failures to determine if the circuit breaker should
+          # open.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.memcached.set-async.circuit-breaker.consecutive-failures
+          [consecutive_failures: <int> | default = 5]
+
+          # Failure percentage to determine if the circuit breaker should open.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.memcached.set-async.circuit-breaker.failure-percent
+          [failure_percent: <float> | default = 0.05]
+
       redis:
         # Comma separated list of redis addresses. Supported prefixes are: dns+
         # (looked up as an A/AAAA query), dnssrv+ (looked up as a SRV query,
@@ -875,6 +962,35 @@ blocks_storage:
         # CLI flag: -blocks-storage.bucket-store.chunks-cache.redis.cache-size
         [cache_size: <int> | default = 0]
 
+        set_async_circuit_breaker_config:
+          # If true, enable circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.redis.set-async.circuit-breaker.enabled
+          [enabled: <boolean> | default = false]
+
+          # Maximum number of requests allowed to pass through when the circuit
+          # breaker is half-open. If set to 0, by default it allows 1 request.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.redis.set-async.circuit-breaker.half-open-max-requests
+          [half_open_max_requests: <int> | default = 10]
+
+          # Period of the open state after which the state of the circuit
+          # breaker becomes half-open. If set to 0, by default open duration is
+          # 60 seconds.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.redis.set-async.circuit-breaker.open-duration
+          [open_duration: <duration> | default = 5s]
+
+          # Minimal requests to trigger the circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.redis.set-async.circuit-breaker.min-requests
+          [min_requests: <int> | default = 50]
+
+          # Consecutive failures to determine if the circuit breaker should
+          # open.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.redis.set-async.circuit-breaker.consecutive-failures
+          [consecutive_failures: <int> | default = 5]
+
+          # Failure percentage to determine if the circuit breaker should open.
+          # CLI flag: -blocks-storage.bucket-store.chunks-cache.redis.set-async.circuit-breaker.failure-percent
+          [failure_percent: <float> | default = 0.05]
+
       # Size of each subrange that bucket object is split into for better
       # caching.
       # CLI flag: -blocks-storage.bucket-store.chunks-cache.subrange-size
@@ -946,6 +1062,35 @@ blocks_storage:
         # CLI flag: -blocks-storage.bucket-store.metadata-cache.memcached.auto-discovery
         [auto_discovery: <boolean> | default = false]
 
+        set_async_circuit_breaker_config:
+          # If true, enable circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.memcached.set-async.circuit-breaker.enabled
+          [enabled: <boolean> | default = false]
+
+          # Maximum number of requests allowed to pass through when the circuit
+          # breaker is half-open. If set to 0, by default it allows 1 request.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.memcached.set-async.circuit-breaker.half-open-max-requests
+          [half_open_max_requests: <int> | default = 10]
+
+          # Period of the open state after which the state of the circuit
+          # breaker becomes half-open. If set to 0, by default open duration is
+          # 60 seconds.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.memcached.set-async.circuit-breaker.open-duration
+          [open_duration: <duration> | default = 5s]
+
+          # Minimal requests to trigger the circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.memcached.set-async.circuit-breaker.min-requests
+          [min_requests: <int> | default = 50]
+
+          # Consecutive failures to determine if the circuit breaker should
+          # open.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.memcached.set-async.circuit-breaker.consecutive-failures
+          [consecutive_failures: <int> | default = 5]
+
+          # Failure percentage to determine if the circuit breaker should open.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.memcached.set-async.circuit-breaker.failure-percent
+          [failure_percent: <float> | default = 0.05]
+
       redis:
         # Comma separated list of redis addresses. Supported prefixes are: dns+
         # (looked up as an A/AAAA query), dnssrv+ (looked up as a SRV query,
@@ -1042,6 +1187,35 @@ blocks_storage:
         # CLI flag: -blocks-storage.bucket-store.metadata-cache.redis.cache-size
         [cache_size: <int> | default = 0]
 
+        set_async_circuit_breaker_config:
+          # If true, enable circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.redis.set-async.circuit-breaker.enabled
+          [enabled: <boolean> | default = false]
+
+          # Maximum number of requests allowed to pass through when the circuit
+          # breaker is half-open. If set to 0, by default it allows 1 request.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.redis.set-async.circuit-breaker.half-open-max-requests
+          [half_open_max_requests: <int> | default = 10]
+
+          # Period of the open state after which the state of the circuit
+          # breaker becomes half-open. If set to 0, by default open duration is
+          # 60 seconds.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.redis.set-async.circuit-breaker.open-duration
+          [open_duration: <duration> | default = 5s]
+
+          # Minimal requests to trigger the circuit breaker.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.redis.set-async.circuit-breaker.min-requests
+          [min_requests: <int> | default = 50]
+
+          # Consecutive failures to determine if the circuit breaker should
+          # open.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.redis.set-async.circuit-breaker.consecutive-failures
+          [consecutive_failures: <int> | default = 5]
+
+          # Failure percentage to determine if the circuit breaker should open.
+          # CLI flag: -blocks-storage.bucket-store.metadata-cache.redis.set-async.circuit-breaker.failure-percent
+          [failure_percent: <float> | default = 0.05]
+
       # How long to cache list of tenants in the bucket.
       # CLI flag: -blocks-storage.bucket-store.metadata-cache.tenants-list-ttl
       [tenants_list_ttl: <duration> | default = 15m]