Temporal server loadtest

Community Support
1

Hi,

We are running a load test against Mysql-based temporal server. I am able to achieve around 1k/s workflow task throughput. With more load, I start to see the context exceed error log piling up in Temporal server when MySQL is at about 50% CPU utilization. I wonder if there is headroom for me to further optimize, or we have hit the capacity limit of MySQL? Thanks!

2

And we are seeting lots of server errors like this

Blockquote
{“level”:“error”,“ts”:“2021-01-26T21:48:49.420Z”,“msg”:“Operation failed with internal error.”,“service”:“history”,“error”:“AppendHistoryEvents: context deadline exceeded”,“metric-scope”:222,“logging-call-at”:“persistenceMetricClients.go:1235”,“stacktrace”:“go.temporal.io/server/common/log/loggerimpl.(*loggerImpl).Error\n\t/temporal/common/log/loggerimpl/logger.go:138\ngo.temporal.io/server/common/persistence.(*historyV2PersistenceClient).updateErrorMetric\n\t/temporal/common/persistence/persistenceMetricClients.go:1235\ngo.temporal.io/server/common/persistence.(*historyV2PersistenceClient).AppendHistoryNodes\n\t/temporal/common/persistence/persistenceMetricClients.go:1134\ngo.temporal.io/server/service/history/shard.(*ContextImpl).AppendHistoryV2Events\n\t/temporal/service/history/shard/context_impl.go:769\ngo.temporal.io/server/service/history.(*workflowExecutionContextImpl).appendHistoryV2EventsWithRetry.func1\n\t/temporal/service/history/workflowExecutionContext.go:945\ngo.temporal.io/server/common/backoff.Retry\n\t/temporal/common/backoff/retry.go:103\ngo.temporal.io/server/service/history.(*workflowExecutionContextImpl).appendHistoryV2EventsWithRetry\n\t/temporal/service/history/workflowExecutionContext.go:949\ngo.temporal.io/server/service/history.(*workflowExecutionContextImpl).persistNonFirstWorkflowEvents\n\t/temporal/service/history/workflowExecutionContext.go:923\ngo.temporal.io/server/service/history.(*workflowExecutionContextImpl).updateWorkflowExecutionWithNew\n\t/temporal/service/history/workflowExecutionContext.go:693\ngo.temporal.io/server/service/history.(*workflowExecutionContextImpl).updateWorkflowExecutionAsActive\n\t/temporal/service/history/workflowExecutionContext.go:601\ngo.temporal.io/server/service/history.(*historyEngineImpl).updateWorkflowHelper\n\t/temporal/service/history/historyEngine.go:2413\ngo.temporal.io/server/service/history.(*historyEngineImpl).updateWorkflow\n\t/temporal/service/history/historyEngine.go:2351\ngo.temporal.io/server/service/history.(*historyEngineImpl).SignalWorkflowExecution\n\t/temporal/service/history/historyEngine.go:1780\ngo.temporal.io/server/service/history.(*Handler).SignalWorkflowExecution\n\t/temporal/service/history/handler.go:956\ngo.temporal.io/server/api/historyservice/v1._HistoryService_SignalWorkflowExecution_Handler.func1\n\t/temporal/api/historyservice/v1/service.pb.go:1073\ngo.temporal.io/server/common/rpc.ServiceErrorInterceptor\n\t/temporal/common/rpc/grpc.go:100\ngo.temporal.io/server/api/historyservice/v1._HistoryService_SignalWorkflowExecution_Handler\n\t/temporal/api/historyservice/v1/service.pb.go:1075\ngoogle.golang.org/grpc.(*Server).processUnaryRPC\n\t/go/pkg/mod/google.golang.org/grpc@v1.33.2/server.go:1210\ngoogle.golang.org/grpc.(*Server).handleStream\n\t/go/pkg/mod/google.golang.org/grpc@v1.33.2/server.go:1533\ngoogle.golang.org/grpc.(*Server).serveStreams.func1.2\n\t/go/pkg/mod/google.golang.org/grpc@v1.33.2/server.go:871”}

3

This is my cluster setup params:

  • env:
    - name: NUM_HISTORY_SHARDS
    value: “1024”
    - name: DB
    value: mysql
    - name: DYNAMIC_CONFIG_FILE_PATH
    value: config/dynamicconfig/development.yaml
    - name: MYSQL_PWD
    valueFrom:
    secretKeyRef:
    key: password
    name: cloudsql-db-credentials
    - name: MYSQL_SEEDS
    value: 127.0.0.1
    - name: MYSQL_USER
    valueFrom:
    secretKeyRef:
    key: username
    name: cloudsql-db-credentials
    - name: SQL_MAX_CONNS
    value: “400”
    - name: SQL_MAX_IDLE_CONNS
    value: “200”
    - name: SQL_MAX_CONN_TIME
    value: 1h
    - name: MYSQL_TX_ISOLATION_COMPAT
    value: “true”
    - name: STATSD_ENDPOINT
    value: 127.0.0.1:8125
    - name: BIND_ON_IP
    value: 0.0.0.0
    - name: TEMPORAL_BROADCAST_ADDRESS
    valueFrom:
    fieldRef:
    apiVersion: v1
    fieldPath: status.podIP
    data:
    dynamic_config.yaml: |-
    frontend.enableClientVersionCheck:
    • value: true
      constraints: {}
      system.minRetentionDays:
    • value: 0
      constraints: {}
      history.persistenceMaxQPS:
    • value: 35000
      constraints: {}
      frontend.persistenceMaxQPS:
    • value: 30000
      constraints: {}
      frontend.historyMgrNumConns:
    • value: 100
      constraints: {}
      frontend.throttledLogRPS:
    • value: 20
      constraints: {}
      history.historyMgrNumConns:
    • value: 500
      constraints: {}
      system.advancedVisibilityWritingMode:
    • value: “off”
      constraints: {}
      matching.numTaskqueueReadPartitions:
    • value: 64
      constraints: {}
      matching.numTaskqueueWritePartitions:
    • value: 64
      constraints: {}
      matching.maxTaskBatchSize:
    • value: 1000
      constraints: {}
      history.defaultActivityRetryPolicy:
    • value:
      InitialIntervalInSeconds: 1
      MaximumIntervalCoefficient: 100.0
      BackoffCoefficient: 2.0
      MaximumAttempts: 0
      history.defaultWorkflowRetryPolicy:
    • value:
      InitialIntervalInSeconds: 1
      MaximumIntervalCoefficient: 100.0
      BackoffCoefficient: 2.0
      MaximumAttempts: 0
4

Can you plz provide the following?

General
MySQL: CPU utilization / CPU load / mem utilization / IO Capacity / IO utilization
Temporal: frontend / matching history server CPU utilization / CPU load / mem utilization

Persistence metrics: persistence_latency
Matching service metrics: poll_success_per_tl and poll_success_sync_per_tl

5

these 2 values maybe too high BTW