Endeca Data Layer Performance Test
Endeca Server is the control center for Endeca data stores. Endeca Server uses a database-like operational model to manage the Endeca data stores running on the machine. The Endeca data stores are created and controlled by the server using a set of commands. Each Endeca data store is serviced by a Dgraph process. The Dgraph uses proprietary data structures and algorithms that allow it to provide real-time responses to client requests. It stores the data files that were created from loading the data into it. After the data files are stored, the Dgraph receives client requests via the application tier, queries the data files, and returns the results. Any slow down in processing the client requests indicate a serious processing bottleneck on the server, probably because Dgraph is taking too long a time to process data files. Administrators should hence continuously track the time taken by the Dgraph to process the data in the data files and return the result to client requests.
This test auto-discovers the data files of the Endeca server, and reports the time taken to flush, commit and merge the data into each data file. Additionally, this test reveals the time taken to release data from each data file and abort any I/O activity on each data file. Using this test, administrators can isolate those data files on which data is taking too long to be flushed/committed.
Target of the test : An Endeca Search Application
Agent deploying the test : An internal agent
Outputs of the test : One set of results for every datafile on the Endeca Search application being monitored.
Parameter | Description |
---|---|
Test period |
How often should the test be executed |
Host |
The IP address of the host for which this test is to be configured. |
Port |
The port at which the specified host listens to. The default port is 15000. |
PerfStatURL |
Specify the URL to access the page from which this test will be collecting the required statistics. The default URL will be in the format: http://{Endeca_Server_IP}:8000/admin?op=stats |
Measurement | Description | Measurement Unit | Interpretation |
---|---|---|---|
Flush time
|
Indicates the time taken to flush the data from this data file. |
Seconds |
The value of this measure depends on how frequently the data is being flushed. Compare the value of this measure across data files to identify the data file on which flushing the data took too long to complete. |
Merge time
|
Indicates the time taken to merge the data in this data file. |
Seconds |
Ideally, the value of this measure should be low. A high value could indicate bottlenecks while merging. Compare the value of this measure across data files to identify the data file on which merging the data took too long to complete. |
Abort time
|
Indicates the time taken to abort a process on this data file. |
Seconds |
The value of this measure should be low. Compare the value of this measure across data files to identify the data file on which aborts took too long to complete. |
Commit time
|
Indicates the time taken to commit the data to this data file. |
Seconds |
Any changes made during the commit process is immediately made permanent. A shorter time duration indicates that the changes are committed at a faster pace. Compare the value of this measure across data files to identify the data file on which data commits took too long to complete. |
Release time |
Indicates the time taken for release the data from this data file. |
Seconds |
The value of this measure should be very low. Compare the value of this measure across data files to identify the data file on which releasing the data took too long to complete. |