Files
ShivamandGitHub 318317fd76 dropping non-retryable activity/wf tasks (#6535)
## What changed?
<!-- Describe what has changed in this PR -->
- This PR introduces changes to drop activity + workflow tasks which are
non-retryable due to certain specific errors.

## Why?
<!-- Tell your future self why have you made these changes -->
- Currently, these tasks are being added back to persistence on failure
and end up wasting resources by continuous retries.
- Manual intervention is required to terminate/cancel these workflows. 

## How did you test it?
<!-- How have you verified this change? Tested locally? Added a unit
test? Checked in staging env? -->
- Added unit tests 
- Existing suite of tests


## Is hotfix candidate?
<!-- Is this PR a hotfix candidate or does it require a notification to
be sent to the broader community? (Yes/No) -->
nope
2024-10-10 10:58:37 -04:00

256 lines
8.2 KiB
Go

// The MIT License
//
// Copyright (c) 2020 Temporal Technologies Inc. All rights reserved.
//
// Copyright (c) 2020 Uber Technologies, Inc.
//
// Permission is hereby granted, free of charge, to any person obtaining a copy
// of this software and associated documentation files (the "Software"), to deal
// in the Software without restriction, including without limitation the rights
// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
// copies of the Software, and to permit persons to whom the Software is
// furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
// THE SOFTWARE.
//go:generate mockgen -copyright_file ../../../LICENSE -package $GOPACKAGE -source $GOFILE -destination events_cache_mock.go
package events
import (
"context"
"time"
historypb "go.temporal.io/api/history/v1"
"go.temporal.io/api/serviceerror"
"go.temporal.io/server/common"
"go.temporal.io/server/common/cache"
"go.temporal.io/server/common/log"
"go.temporal.io/server/common/log/tag"
"go.temporal.io/server/common/metrics"
"go.temporal.io/server/common/namespace"
"go.temporal.io/server/common/persistence"
"go.temporal.io/server/common/persistence/serialization"
"go.temporal.io/server/service/history/configs"
)
type (
EventKey struct {
NamespaceID namespace.ID
WorkflowID string
RunID string
EventID int64
Version int64
}
Cache interface {
GetEvent(ctx context.Context, shardID int32, key EventKey, firstEventID int64, branchToken []byte) (*historypb.HistoryEvent, error)
PutEvent(key EventKey, event *historypb.HistoryEvent)
DeleteEvent(key EventKey)
}
CacheImpl struct {
cache.Cache
executionManager persistence.ExecutionManager
metricsHandler metrics.Handler
logger log.Logger
disabled bool
}
historyEventCacheItemImpl struct {
event *historypb.HistoryEvent
}
)
var (
errEventNotFoundInBatch = serviceerror.NewInternal("History event not found within expected batch")
)
var _ Cache = (*CacheImpl)(nil)
func NewHostLevelEventsCache(
executionManager persistence.ExecutionManager,
config *configs.Config,
handler metrics.Handler,
logger log.Logger,
disabled bool,
) Cache {
return newEventsCache(executionManager, handler, logger, config.EventsHostLevelCacheMaxSizeBytes(), config.EventsCacheTTL(), disabled)
}
func NewShardLevelEventsCache(
executionManager persistence.ExecutionManager,
config *configs.Config,
handler metrics.Handler,
logger log.Logger,
disabled bool,
) Cache {
return newEventsCache(executionManager, handler, logger, config.EventsShardLevelCacheMaxSizeBytes(), config.EventsCacheTTL(), disabled)
}
func newEventsCache(
executionManager persistence.ExecutionManager,
metricsHandler metrics.Handler,
logger log.Logger,
maxSize int,
ttl time.Duration,
disabled bool,
) *CacheImpl {
opts := &cache.Options{}
opts.TTL = ttl
taggedMetricHandler := metricsHandler.WithTags(metrics.CacheTypeTag(metrics.EventsCacheTypeTagValue))
return &CacheImpl{
Cache: cache.NewWithMetrics(maxSize, opts, taggedMetricHandler),
executionManager: executionManager,
metricsHandler: taggedMetricHandler,
logger: logger,
disabled: disabled,
}
}
func (e *CacheImpl) validateKey(key EventKey) bool {
if len(key.NamespaceID) == 0 || len(key.WorkflowID) == 0 || len(key.RunID) == 0 || key.EventID < common.FirstEventID {
// This is definitely a bug, but just warn and don't crash so we can find anywhere this happens.
e.logger.Warn("one or more ids is invalid in event cache",
tag.WorkflowID(key.WorkflowID),
tag.WorkflowRunID(key.RunID),
tag.WorkflowNamespaceID(key.NamespaceID.String()),
tag.WorkflowEventID(key.EventID))
return false
}
return true
}
func (e *CacheImpl) GetEvent(ctx context.Context, shardID int32, key EventKey, firstEventID int64, branchToken []byte) (*historypb.HistoryEvent, error) {
handler := e.metricsHandler.WithTags(metrics.OperationTag(metrics.EventsCacheGetEventScope), metrics.NamespaceIDTag(key.NamespaceID.String()))
metrics.CacheRequests.With(handler).Record(1)
startTime := time.Now().UTC()
defer func() { metrics.CacheLatency.With(handler).Record(time.Since(startTime)) }()
validKey := e.validateKey(key)
// Test hook for disabling cache
if !e.disabled {
eventItem, cacheHit := e.Cache.Get(key).(*historyEventCacheItemImpl)
if cacheHit {
return eventItem.event, nil
}
}
metrics.CacheMissCounter.With(handler).Record(1)
event, err := e.getHistoryEventFromStore(ctx, shardID, key, firstEventID, branchToken)
if err != nil {
metrics.CacheFailures.With(handler).Record(1)
e.logger.Error("Cache unable to retrieve event from store",
tag.Error(err),
tag.WorkflowID(key.WorkflowID),
tag.WorkflowRunID(key.RunID),
tag.WorkflowNamespaceID(key.NamespaceID.String()),
tag.WorkflowEventID(key.EventID))
return nil, err
}
// If invalid, return event anyway, but don't store in cache
if validKey {
e.put(key, event)
}
return event, nil
}
func (e *CacheImpl) PutEvent(key EventKey, event *historypb.HistoryEvent) {
handler := e.metricsHandler.WithTags(metrics.OperationTag(metrics.EventsCachePutEventScope), metrics.NamespaceIDTag(key.NamespaceID.String()))
metrics.CacheRequests.With(handler).Record(1)
startTime := time.Now().UTC()
defer func() { metrics.CacheLatency.With(handler).Record(time.Since(startTime)) }()
if !e.validateKey(key) {
return
}
e.put(key, event)
}
func (e *CacheImpl) DeleteEvent(key EventKey) {
handler := e.metricsHandler.WithTags(metrics.OperationTag(metrics.EventsCacheDeleteEventScope), metrics.NamespaceIDTag(key.NamespaceID.String()))
metrics.CacheRequests.With(handler).Record(1)
startTime := time.Now().UTC()
defer func() { metrics.CacheLatency.With(handler).Record(time.Since(startTime)) }()
e.validateKey(key) // just for log message, delete anyway
e.Delete(key)
}
func (e *CacheImpl) getHistoryEventFromStore(
ctx context.Context,
shardID int32,
key EventKey,
firstEventID int64,
branchToken []byte,
) (*historypb.HistoryEvent, error) {
handler := e.metricsHandler.WithTags(metrics.OperationTag(metrics.EventsCacheGetFromStoreScope), metrics.NamespaceIDTag(key.NamespaceID.String()))
metrics.CacheRequests.With(handler).Record(1)
startTime := time.Now().UTC()
defer func() { metrics.CacheLatency.With(handler).Record(time.Since(startTime)) }()
response, err := e.executionManager.ReadHistoryBranch(ctx, &persistence.ReadHistoryBranchRequest{
BranchToken: branchToken,
MinEventID: firstEventID,
MaxEventID: key.EventID + 1,
PageSize: 1,
NextPageToken: nil,
ShardID: shardID,
})
switch err.(type) {
case nil:
// noop
case *serviceerror.DataLoss, *serialization.DeserializationError, *serialization.SerializationError:
// log event
e.logger.Error("encounter data corruption event",
tag.WorkflowNamespaceID(key.NamespaceID.String()),
tag.WorkflowID(key.WorkflowID),
tag.WorkflowRunID(key.RunID))
metrics.CacheFailures.With(handler).Record(1)
return nil, err
default:
metrics.CacheFailures.With(handler).Record(1)
return nil, err
}
// find history event from batch and return back single event to caller
for _, e := range response.HistoryEvents {
if e.EventId == key.EventID && e.Version == key.Version {
return e, nil
}
}
return nil, errEventNotFoundInBatch
}
func (e *CacheImpl) put(key EventKey, event *historypb.HistoryEvent) interface{} {
return e.Put(key, newHistoryEventCacheItem(event))
}
var _ cache.SizeGetter = (*historyEventCacheItemImpl)(nil)
func newHistoryEventCacheItem(
event *historypb.HistoryEvent,
) *historyEventCacheItemImpl {
return &historyEventCacheItemImpl{
event: event,
}
}
func (h *historyEventCacheItemImpl) CacheSize() int {
return h.event.Size()
}