stats_instance()
The stats_instance() operator provides per-instance resource statistics. Available only in the Enterprise Edition.
Synopsis
stats_instance()
Provides per-instance resource statistics.
Library
The stats_instance() operator resides in the monitor library. Run the following query to load this library:
Â
AFL% load_library('monitor');
Summary
The stats_instance() operator lets you gather per-instance SciDB resource statistics for diagnostic purposes. Â
Attributes
Â
Attribute | Description |
---|---|
host_name | Node's hostname |
ip_address | Node's Internet Protocol address |
process_id | The SciDB process ID |
time_since_start_ms | Time elapsed since the instance started in milliseconds |
time_since_reset_ms | Time since reset occurred (or instance start) in milliseconds |
time_cpu_active_ms | Total CPU active time since the report was reset (or instance start) in milliseconds |
net_send_bytes | Size of chunks sent across the network in bytes |
net_recv_bytes | Size of chunks received from the network in bytes |
mem_available_bytes | Total size of all RAM currently available in bytes |
mem_allocated_bytes | RAM allocated on behalf of the instance (at this moment) in bytes |
mem_allocations | Number of times RAM was allocated on this instance in bytes |
mem_peak_usage_bytes | Maximum amount of RAM allocated on this instance in bytes |
num_active_queries | Number of active queries currently running on the instance |
cache_used_mem_size | Total size of all mem chunks currently in cache in bytes |
cache_swap_bytes | Size of mem chunks written to disk (cache swap out) in bytes |
cache_swap_num | Number of mem chuncks written to disk |
cache_load_bytes | Size of mem chunks read off disk (due to read cache miss) in bytes |
cache_load_num | Number of mem chunks read off disk |
cache_drop_bytes | Size of mem chunks dropped from cache in bytes |
cache_drop_num | Number of mem chunks dropped from cache |
time_cache_write_ms | Time spent waiting on Mem chunk writes in milliseconds |
time_cache_read_ms | Time spent waiting on Mem chunk reads in milliseconds |
num_db_wr | Number of DB Chunks Written |
bytes_compr_db_wr | Total size of all compressed DB Chunks written to disk in bytes |
bytes_decomp_db_wr | Total size of all decompressed DB Chunks written to disk in bytes |
num_db_rd | Number of compressed DB Chunks read from disk |
bytes_compr_db_rd | Total size of all compressed DB Chunks read from disk in bytes |
bytes_decomp_db_rd | Total size of all decompressed DB Chunks read from disk in bytes |
time_catalog_access_ms | Time spent waiting on the execution of system catalogue queries in milliseconds |
Some of the attributes are disabled if the resource-monitoring variable set to False in the config.ini.
Examples
Full stats_instance() Statistics
To create the full set of stats_instance() statistics, run the following:
$ iquery -aq "stats_instance();"
The output is:
{inst,n} host_name,ip_address,process_id,time_since_start_ms,time_since_reset_ms,time_cpu_active_ms,net_send_bytes,net_recv_bytes,mem_available_bytes,mem_allocated_bytes,mem_allocations,mem_peak_usage_bytes,num_active_queries,cache_used_mem_size,cache_swap_bytes,cache_swap_num,cache_load_bytes,cache_load_num,cache_drop_bytes,cache_drop_num,time_cache_write_ms,time_cache_read_ms,num_db_wr,bytes_compr_db_wr,bytes_decomp_db_wr,num_db_rd,bytes_compr_db_rd,bytes_decomp_db_rd,time_catalog_access_ms {0,0} 'goofy.example.com','0.0.0.0',13174,30522,30522,20,1478,7482,1152921504606846975,67117080,1,201334840,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4 {1,0} 'goofy.example.com','0.0.0.0',13173,30523,30523,11,432,9390,1152921504606846975,67117080,1,67117080,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,3 {2,0} 'goofy.example.com','0.0.0.0',13177,30521,30521,8,432,9404,1152921504606846975,67117080,1,67117080,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2 {3,0} 'goofy.example.com','0.0.0.0',13175,30522,30522,9,432,9390,1152921504606846975,67117080,1,67117080,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4
Subset of stats_instance() Statistics
To create a subset of stats_instance() statistics (this example captures memory statistics), run the following:
$ iquery -aq "project(stats_instance(), mem_available_bytes, mem_allocations, mem_allocated_bytes, mem_peak_usage_bytes);"
The output is:
{inst,n} mem_available_bytes,mem_allocations,mem_allocated_bytes,mem_peak_usage_bytes {0,0} 1152921504606846975,1,67117080,201334840 {1,0} 1152921504606846975,1,67117080,67117080 {2,0} 1152921504606846975,1,67117080,67117080 {3,0} 1152921504606846975,1,67117080,67117080
Individual stats_instance() Statistic
To view an individual stats_instance() statistic (in this case, the size of chunks sent across the network in bytes), run the following:
$ iquery -aq "project(stats_instance(), net_send_bytes);"
The output is:
{inst,n} net_send_bytes {0,0} 7004 {1,0} 2160 {2,0} 2160 {3,0} 2160