Monitor the TiFlash Cluster
This document describes the monitoring items of TiFlash.
Monitor the Coprocessor
|The number of coprocessor requests received. |
batch is the number of batch requests.
batch_cop is the number of coprocessor requests in the batch requests.
cop is the number of coprocessor requests that are sent directly via the coprocessor interface.
cop_dag is the number of dag requests in all coprocessor requests.
|The number of each type of dag executors. |
table_scan is the table scan executor.
selection is the selection executor.
aggregation is the aggregation executor.
top_n is the
limit is the limit executor.
|The histogram of the duration of each coprocessor request, in which the duration is from the time that the coprocessor request is received to the time that the response to the request is completed. |
batch is the duration of batch requests.
cop is the duration of coprocessor requests that are sent directly via the coprocessor interface.
|The number of errors of coprocessor requests. |
meet_lock means that the read data is locked.
region_not_found means that the Region does not exist.
epoch_not_match means the read Region epoch is inconsistent with the local epoch.
kv_client_error means that the communication with TiKV returns an error.
internal_error is the internal system error of TiFlash.
other is other type of errors.
|The histogram of the processing time of each coprocessor request, in which the processing time is from starting to execute the coprocessor request to completing the execution. |
batch is the processing time of batch request.
cop is the processing time of coprocessor requests that are sent directly via the coprocessor interface.
|The total bytes of the response.|
Monitor DDL operations
|The version of the schema currently cached in TiFlash.|
|This item includes the count of three types of |
full apply, and
diff apply is the normal process of a single apply. If
diff apply fails,
failed apply increases by
1, and TiFlash rolls back to
|The number of specific DDL operations in TiFlash.|
|The time used for a single |
apply schema operation.
|The number of times that the coprocessor triggers the |
read_index request, which equals to the number of Regions triggered by a coprocessor.
|The time used by |
read_index. Most time is used for interaction with Leader and retry.
|The time used by |
wait_index, namely the time used to wait until local index >= read_index after the
read_index request is received.