Search Gatherer Transactions Test

Crawls, when scheduled to occur too frequently, can significantly impact the processing ability of the content processing component, the level of I/O activity on the crawl database, and ultimately, the search throughput! Likewise, a resource-starved content processing component and/or a crawl database can also considerably slowdown SharePoint search, as they may not be able to handle the workload generated by the crawler! This is why, when end-users complain of slow searching by SharePoint, administrators need to be able to quickly figure out where the bottleneck is and how to clear it – should the crawl schedules be changed so that less crawls occur? Or should the processing power of the content processor and crawl database change in tandem with the frequency of crawls? This is where the Search Gatherer Transactions test helps!

This test monitors the transactions on the crawl component and reports the count of transactions that are waiting for processing by the content processor and those that have completed processing. In the process, the test turns the spotlight on a potential processing slowdown and accurately pinpoints what is causing it – is it owing to too many crawls? Or is it because the content processor and/or the crawl database are incorrectly sized? Based on the results of this test, administrators can clearly understand what needs to be fine-tuned and how.

Note:

This test will report metrics only if the crawl component is available on the target SharePoint server.

Target of the test : A Microsoft SharePoint Server

Agent deploying the test : An internal agent

Outputs of the test : One set of results each for the Microsoft SharePoint server that is being monitored

Configurable parameters for the test
Parameters Description

Test period

This indicates how often should the test be executed.

Host

The host for which the test is to be configured.

Port

The port at which the host server listens.

Measurements made by the test
Measurement Description Measurement Unit Interpretation

Waiting transactions

Indicates the number of transactions that are currently waiting to be processed by the content processing component.

Number

Ideally, this value should be low (less than a few thousand). If so, it implies that content processing is keeping up with content crawling.

On the other hand, if the value of this measure is high  and/or consistently rising, then it means that the crawl component is pushing more data for processing than what the content processing component can handle. This will slow down content processing and eventually affect SharePoint search! Under such circumstances, you can do either of the following:

  • Provide more processing power to the content processing component, so that it is able to handle the load imposed by the crawl component. You can also add more content processing components to uniformly distribute the processing load.
  • Reconfigure the crawl component to run crawls less frequently, so that the crawl component does not overload the content processing component

Transactions in progress

Indicates the number of transactions that are currently being processed by the crawl component.

Number

This is a good indicator of the current load on the crawl component.

Completed transactions

Indicates the number of transactions that are completed

Number

If this value is very high (say, greater than a few hundred), it means that too many transactions are getting completed and are written to the crawl database, causing disk activity on the database to increase. At this juncture, check the crawl database for disk latency. If the disk latency and disk queue length are also high, you can conclude that the crawl database is where the bottleneck is.