[jira] [Commented] (IGNITE-8078) Add new metrics for data storage

["Vladimir Ozerov (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IGNITE-8078?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16420513#comment-16420513 ] 

Vladimir Ozerov commented on IGNITE-8078:
-----------------------------------------

[~DmitriyGovorukhin], what is the purpose of {{IndexMxBean.getProcessedPuts}} and {{IndexMxBean.getProcessedGets}} metrics? Do not make sense to me, because typically we do not do "gets" on indexes, we do searches instead. And for "puts" it is not clear how to align them with dynamic index creation. I would avoid exposing this stuff to public API. Instead, we'd better to expose index statistics here (histograms, selectivity, clustering, etc) when they are ready.

> Add new metrics for data storage
> --------------------------------
>
>                 Key: IGNITE-8078
>                 URL: https://issues.apache.org/jira/browse/IGNITE-8078
>             Project: Ignite
>          Issue Type: New Feature
>            Reporter: Dmitriy Govorukhin
>            Priority: Major
>              Labels: iep-6
>
> 1. Create new MXbean for each index, IndexMxBean
> {code}
> class IndexMxBean{
> /** The number of PUT operations on the index. */
> long getProcessedPuts();
> /** The number of GET operations on the index. */
> long getProcessedGets();
> /** The total index size in bytes. */
> long getIndexSize();
> /** Index name.*/
> String getName();
> }
> {code}
> 2. Add new metrics for data storage and cache group.
> {code}
> class CacheGroupMetricsMXBean{
> /** The total index size in bytes */
> long getIndexesSize();
> /** Total size in bytes for primary key indexes. */
> long getPKIndexesSize();
> /** Total size in bytes. */
> long getTotalSize();
> /** Total size in bytes for pure data.*/
> long getDataSize();
> /** CacheGroup type. PARTITIONED, REPLICATED, LOCAL.*/
> String getType();
> /** Partitions currently assigned to the local node in this cache group. */
> int[] getPartitions();
> }
> {code}
> {code}
> class DataRegionMXBean{
> /** Total size in bytes for indexes. */
> long getIndexesSize();
> /** Total size in bytes for primary key indexes. */
> long getPKIndexesSize();
> /** Total size in bytes. */
> long getTotalSize();
> /** Total size in bytes for pure data.*/
> long getDataSize();
> /** Total used offheap size in bytes. */
> long getOffheapUsedSize();
> /** The number of read pages from last restart. */
> long getPagesRead();
> /** The number of writen pages from last restart. */
> long getPagesWriten();
> /** The number of replaced pages from last restart . */
> long getPagesReplaced();
> /** Total dirty pages for the next checkpoint. */
> long getDirtyPagesForNextCheckpoint();
> }
> {code}
> {code}
> class DataStorageMXbean{
> /** Total size in bytes for indexes. */
> long getIndexesSize();
> /** Total size in bytes for primary key indexes. */
> long getPKIndexesSize();
> /** Total size in bytes for all storages. */
> long getTotalSize();
> /** Total offheap size in bytes. */
> long getOffHeapSize();
> /** Total used offheap size in bytes for all data regions. */
> long getOffheapUsedSize();
> /** Total size in bytes for pure data.*/
> long getDataSize();
> /** The number of read pages from last restart. */
> long getPagesRead();
> /** The number of writen pages from last restart. */
> long getPagesWriten();
> /** The number of replaced pages from last restart. */
> long getPagesReplaced();
> /** Total checkpoint time from last restart. */
> long getCheckpointTotalTime();
> /** Total dirty pages for the next checkpoint. */
> long getDirtyPagesForNextCheckpoint();
> /** Total size in bytes for storage wal files. */
> long getWalTotalSize();
> /** Time of the last WAL segment rollover. */
> long getWalLastSwitchTime();
> }
> {code}
> {code}
> class IgniteMxBean {
> /** Returns string containing Node ID, Consistent ID, Node Order */
> String getCurrentCoordinator();
> }
> {code}
> Depricate CacheMetrics.getRebalancingPartitionsCount(); and move to CacheGroupMetricsMXBean.getRebalancingPartitionsCount();



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)