Yandex Managed Service for YDB metrics

Written by

Yandex Cloud

Improved by

Max Z.

Updated at April 28, 2025

Database metrics
Resource usage metrics
API metrics
Session metrics
Transaction processing metrics
Query processing metrics
Table partition metrics (DataShards)
Table partition metrics (ColumnShards)
Resource usage metrics (for Dedicated mode only)
Query processing metrics (for Dedicated mode only)

This section describes Managed Service for YDB metrics delivered to Monitoring.

The name label stores the metric name.

Database metrics

Metric name Type, units	Description Labels
`database_size` `DGAUGE`, bytes	Database size

Resource usage metrics

Metric name Type, units	Description Labels
`resources.storage.limit_bytes` `IGAUGE`, bytes	Limit on the size of user and system data a database can store in a distributed network storage.
`resources.storage.topic.used_bytes` `DGAUGE`, bytes	Topic storage size in use
`resources.storage.used_bytes` `IGAUGE`, bytes	Size of user and system data stored in the distributed network storage. System data includes primary and secondary index data.
`resources.stream.limit_shards` `DGAUGE`, count	Limit on the number of shards per stream
`resources.stream.storage.limit_bytes` `DGAUGE`, bytes	Stream storage size limit
`resources.stream.storage.reserved_bytes` `DGAUGE`, bytes	Reserved stream storage size
`resources.stream.throughput.limit_bytes_per_second` `DGAUGE`, bytes per second	Stream throughput limit
`resources.stream.used_shards` `DGAUGE`, count	Number of shards used by the stream

API metrics

Metric name Type, units	Description Labels
`api.grpc.request.bytes` `RATE`, bytes	Size of requests received by the database over a certain period of time. Labels: `api_service`: gRPC API service name, e.g., `table`. `method`: gRPC API service method name, e.g., `ExecuteDataQuery`.
`api.grpc.request.count` `RATE`, count	Total DB requests. Labels: `api_service`: gRPC API service name, e.g., `table`. `method`: gRPC API service method name, e.g., `ExecuteDataQuery`.
`api.grpc.request.dropped_count` `RATE`, count	Number of requests dropped at the transport (gRPC) layer due to an error. Labels: `api_service`: gRPC API service name, e.g., `table`. `method`: gRPC API service method name, e.g., `ExecuteDataQuery`.
`api.grpc.request.inflight_bytes` `IGAUGE`, bytes	Size of requests concurrently handled by the database over a certain period of time. Labels: `api_service`: gRPC API service name, e.g., `table`. `method`: gRPC API service method name, e.g., `ExecuteDataQuery`.
`api.grpc.request.inflight_count` `IGAUGE`, count	Number of requests concurrently handled by the database over a certain period of time. Labels: `api_service`: gRPC API service name, e.g., `table`. `method`: gRPC API service method name, e.g., `ExecuteDataQuery`.
`api.grpc.response.bytes` `RATE`, bytes	Size of responses sent by the database over a certain period of time. Labels: `api_service`: gRPC API service name, e.g., `table`. `method`: gRPC API service method name, e.g., `ExecuteDataQuery`.
`api.grpc.response.count` `RATE`, count	Number of responses sent by the database over a certain period of time. Labels: `api_service`: gRPC API service name, e.g., `table`. `method`: gRPC API service method name, e.g., `ExecuteDataQuery`. `status`: Query execution status. For a detailed description of statuses, see Error Handling.
`api.grpc.response.dropped_count` `RATE`, count	Number of responses dropped at the transport (gRPC) layer due to an error. Labels: `api_service`: gRPC API service name, e.g., `table`. `method`: gRPC API service method name, e.g., `ExecuteDataQuery`.
`api.grpc.response.issues` `RATE`, count	Number of specific error types encountered in gRPC API responses over a specified period of time. Labels: `issue_type`: Error type. The only value is `optimistic_locks_invalidation`. For more information about lock invalidation, see YDB transactions and queries.
`api.request.completed_per_second` `DGAUGE`, requests per second	API request completion rate
`api.request.latency_milliseconds` `IGAUGE`, milliseconds	API latency
`api.request.latency_milliseconds_count` `COUNTER`, request count	Total API requests with measured latency
`api.request.latency_milliseconds_sum` `COUNTER`, milliseconds	Overall API latency
`api.request.size_bytes_per_second` `DGAUGE`, bytes per second	API request processing rate
`api.response.size_bytes_per_second` `DGAUGE`, bytes per second	API response processing rate
`api.units.consumed_by_method_per_second` `DGAUGE`, units per second	Rate of resource consumption by a specific API method
`api.units.consumed_per_second` `DGAUGE`, units per second	Overall rate of resource consumption by API methods

Session metrics

Metric name Type, units	Description Labels
`table.session.active_count` `IGAUGE`, count	Number of active client sessions
`table.session.closed_by_idle_count` `RATE`, count	Number of sessions closed by the DB server within a specific time period due to exceeding the idle session timeout

Transaction processing metrics

You can analyze transaction duration using a histogram counter. The intervals are set in milliseconds. The chart shows the number of transactions binned by duration ranges.

Metric name Type, units	Description Labels
`table.transaction.client_duration_milliseconds` `HIST_RATE`, count	Number of client-side transactions of a certain duration. The duration is the client-side wait time between sending individual requests within a single transaction. It does not include the time it takes for the server to process those requests. Labels: `tx_kind`: Transaction type. Possible values are `read_only`, `read_write`, `write_only`, or `pure`.
`table.transaction.server_duration_milliseconds` `HIST_RATE`, count	Number of server-side transaction of a certain duration. The duration is the time it takes for the server to process requests within a transaction. It does not include the client-side wait time between sending individual requests within a single transaction. Labels: `tx_kind`: Transaction type. Possible values are `read_only`, `read_write`, `write_only`, or `pure`.
`table.transaction.total_duration_milliseconds` `HIST_RATE`, count	Number of transactions of a certain duration on both the server and the client. The duration of a transaction is the time interval from the transaction start (explicit or implicit) to either commit or rollback. It includes the server-side transaction processing time and the client-side wait time between sending different requests within a single transaction. Labels: `tx_kind`: Transaction type. Possible values are `read_only`, `read_write`, `write_only`, or `pure`.

Query processing metrics

Metric name Type, units	Description Labels
`table.query.compilation.latency_milliseconds` `HIST_RATE`, count	Histogram counter. The intervals are set in milliseconds. It shows the number of successful table query compilations binned by latency ranges.
`table.query.compilation.active_count` `IGAUGE`, count	Number of active compilations
`table.query.compilation.count` `RATE`, count	Number of compilations completed successfully over a certain time period
`table.query.compilation.error_count` `RATE`, count	Number of compilations failed over a certain time period
`table.query.compilation.cache_hits` `RATE`, count	Number of queries over a certain period of time with no compilation needed due to a pre-existing plan in the prepared query cache
`table.query.compilation.cache_misses` `RATE`, count	Number of queries requiring compilation over a certain period of time
`table.query.execution.latency_milliseconds` `HIST_RATE`, count	Histogram counter. The intervals are set in milliseconds. It shows the number of queries binned by execution time ranges.
`table.query.request.bytes` `RATE`, bytes	Size of YQL query strings and parameter values for queries that entered the database over a certain period of time.
`table.query.request.parameters_bytes` `RATE`, bytes	Size of parameters for queries that entered the database database over a certain period of time.
`table.query.response.bytes` `RATE`, bytes	Size of responses sent by the database over a certain period of time

Table partition metrics (DataShards)

Metric name Type, unit	Description Labels
`table.datashard.bulk_upsert.bytes` `RATE`, bytes	Size of data added through the `BulkUpsert` gRPC API call to all partitions of all DB tables over a certain period of time
`table.datashard.bulk_upsert.rows` `RATE`, count	Number of rows added through the `BulkUpsert` gRPC API call to all partitions of all DB tables over a certain period of time
`table.datashard.erase.bytes` `RATE`, bytes	Size of data deleted from the database over a certain period of time
`table.datashard.erase.rows` `RATE`, count	Number of rows deleted from the database over a certain period of time
`table.datashard.read.bytes` `RATE`, bytes	Size of data read by all partitions of all DB tables over a certain period of time
`table.datashard.read.rows` `RATE`, count	Number of rows read by all partitions of all DB tables over a certain period of time
`table.datashard.row_count` `GAUGE`, count	Number of rows in DB tables
`table.datashard.scan.bytes` `RATE`, bytes	Size of data read through the `StreamExecuteScanQuery` or `StreamReadTable` gRPC API calls by all partitions of all DB tables over a certain period of time
`table.datashard.scan.rows` `RATE`, count	Number of rows read through the `StreamExecuteScanQuery` or `StreamReadTable` gRPC API calls by all partitions of all DB tables over a certain period of time
`table.datashard.size_bytes` `GAUGE`, bytes	Size of data in DB tables
`table.datashard.used_core_percents` `HIST_GAUGE`, %	Histogram counter. The intervals are set as a percentage. It shows the number of table partitions binned by computing resource usage percentage.
`table.datashard.write.rows` `RATE`, count	Number of rows written by all partitions of all DB tables over a certain period of time
`table.datashard.write.bytes` `RATE`, bytes	Size of data written by all partitions of all DB tables over a certain period of time

Table partition metrics (ColumnShards)

Metric name Type, unit	Description Labels
`table.columnshard.bulk_upsert.bytes` `RATE`, bytes per second	Size of data added through the `BulkUpsert` gRPC API call to all partitions of all DB tables over a certain period of time
`table.columnshard.bulk_upsert.rows` `RATE`, rows per second	Number of rows added through the `BulkUpsert` gRPC API call to all partitions of all DB tables over a certain period of time
`table.columnshard.scan.bytes` `RATE`, bytes per second	Size of data read through the `StreamExecuteScanQuery` or `StreamReadTable` gRPC API calls by all partitions of all DB tables over a certain period of time
`table.columnshard.scan.rows` `RATE`, rows per second	Number of rows read through the `StreamExecuteScanQuery` or `StreamReadTable` gRPC API calls by all partitions of all DB tables over a certain period of time

Resource usage metrics (for Dedicated mode only)

Metric name Type units	Description Labels
`resources.cpu.limit_core_percents` `IGAUGE`, %	Percentage of CPU available to a database. For example, for a database that has three nodes with four cores in `pool=user` per node, the value of this metric will be `1200`. Labels: `pool`: Computing pool. Possible values are `user`, `system`, `batch`, `io`, or `ic`.
`resources.cpu.used_core_percents` `RATE`, %	CPU usage. If the value is `100`, one of the cores is being 100% used. The value may be greater than `100` for multi-core configurations. Labels: `pool`: Computing pool. Possible values are `user`, `system`, `batch`, `io`, or `ic`.
`resources.memory.limit_bytes` `IGAUGE`, bytes	RAM available to database nodes
`resources.memory.used_bytes` `IGAUGE`, bytes	RAM used by database nodes

Query processing metrics (for Dedicated mode only)

Metric name Type units	Description Labels
`table.query.compilation.cache_evictions` `RATE`, count	Number of queries evicted from the prepared query cache over a certain period of time
`table.query.compilation.cache_size_bytes` `IGAUGE`, bytes	Prepared query cache size (memory used)
`table.query.compilation.cached_query_count` `IGAUGE`, count	Prepared query cache size (compiled query plans)