Skip to main content

PingCAP Clinic Health Rule Release Note

v1.0.0​

Release Date: July 15, 2022

Metrics Checking Rule

  • [Added] Region size is too large
  • [Added] GC process is abnormal
  • [Added] The CPU of TiDB nodes is not balanced
  • [Added] The memory of TiKV nodes is not balanced
  • [Added] Too many failed QPS
  • [Added] StmtPrepare usage is low
  • [Added] Raft-engine is not enable
  • [Added] Scatter scheduler should be disabled
  • [Added] TiDB CPU is high
  • [Added] TiKV disk usage is high
  • [Added] QPS is imbalance from applications
  • [Added] QPS downward trend follows active connection counts
  • [Added] Plan cache hit rate is low
  • [Added] Latency bottleneck outside TiDB
  • [Added] JDBC set recommended
  • [Added] Active connection counts is not balanced
  • [Added] TiKV CPU is high
  • [Added] TiKV AsyncIO is not enabled
  • [Added] Network latency high between TiDB and TiKV
  • [Added] The CPU of TiKV nodes is not balanced
  • [Added] Active connection increased while QPS not
  • [Added] The latency of TiDB nodes is not balanced
  • [Added] The memory of TiDB nodes is not balanced
  • [Added] PD schedulers are abnormal
  • [Added] Too many heartbeat messages
  • [Added] Get-token is slow in TiDB
  • [Added] Parsing query string is slow in TiDB
  • [Added] Compiling query statement is slow in TiDB
  • [Added] TiDB takes a long time to wait for tso
  • [Added] tso rpc duration is high
  • [Added] Pessimistic lock wait time is too long
  • [Added] The wait duration of raft propose is high
  • [Added] Too many MVCC versions
  • [Added] TiDB coprocessor duration is high

Log Checking Rule

  • [Added] Writing conflict in optimistic transactions
  • [Added] GC life time is shorter than transaction duration
  • [Added] Transaction not commit for a long time
  • [Added] Data index may be inconsistent
  • [Added] Writing conflict in optimistic transactions
  • [Added] RocksDB write stall
  • [Added] TiKV can't bootstrap since region error
  • [Added] Tasks pile up in the scheduler
  • [Added] Pump has insufficient disk space
  • [Added] Pump fails to notify all living drainer
  • [Added] DM sync error with incorrect utf8 value
  • [Added] Lightning checksum error
  • [Added] Timestamp error on tidb lightning
  • [Added] Lightning can't guess encoding for input file
  • [Added] Opened engine files exceed the limit of tikv-importer
  • [Added] Metadata is written too frequently for TiCDC
  • [Added] CDC sync interrupted since buffer limit
  • [Added] Canal requires old value to be enabled
  • [Added] Creating changefeed failed
  • [Added] Requests to TiKV coprocessor queued more than 60s
  • [Added] TiDB send request timeout
  • [Added] TiKV response timeout
  • [Added] TiKV peer is not leader
  • [Added] DM sync abnormal
  • [Added] DM sync interrupted