Managed Solr Performance Monitoring


Overview

SearchStax® Managed Solr provides real-time and historical statistical graphs of critical Solr server and JVM metrics and statistics. These statistics provide a birds-eye view of Solr cluster health, enabling you to tune server performance.

Pulse Data Retention

“Silver” (self-service) accounts retain Pulse data for 24 hours.

Premium accounts (Gold, Platinum, Platinum Plus) generally retain Pulse data for 30 or 90 days, depending on the contract.

If the Pulse graphs appear empty, see Pulse says “No Data,” “Data not available,” or “Downloading data failed.”.

Contents of this page:

How to Access the Pulse Graphs

You can access the graphs from the Managed Solr Dashboard.

To access the graphs:

  • From the Managed Solr Dashboard, select Pulse.
  • Using the droplist in the upper left corner, select a Deployment.
  • Set a time interval from the droplist in the upper right. (The image below shows a “custom” time interval.) Note that short intervals display more quickly than long intervals.
  • Check the servers you want to inspect from the droplist of servers.

The Pulse graphs display data from a starting date/time to an ending date/time. There are several predefined intervals and a “custom” control.

Note the date and time controls at the top of the screen. These let you examine data from a small time interval that occurred days or weeks in the past.

About the Graphs

Managed Solr currently provides graphs for the following server and search metrics.

CPU, Memory, JVM

These are monitoring graphs for a server that is running out of memory. Pay close attention to the Y-axis labels. Pulse zooms in on the data, so the Y-axis is dynamic.

SearchStax Solr Pulse monitoring

The following server-performance metrics are displayed as graphs by Pulse:

MetricDefinition
UsageCPU load as a percent of maximum. (See Is 100% CPU a bad thing?)
System Load AverageNumber of processes per CPU over one-minute intervals. Values below 1.0 indicate that processes are not waiting.
MemoryTotal memory size and used memory size plotted on a single graph. (See How do I recognize a Solr Out-of-Memory Error?)
Swap UsageTotal disc swap space available and used over time.
JVM Heap Memory UsageHeap memory usage over time. Heap memory is used for storing Java objects. When heap memory reaches 100%, Solr stops running.
Non-Heap MemoryNon-heap memory usage over time. Non-heap memory is used by Java to store loaded classes and other metadata.
JVM Thread CountServer thread count usage over time.
1 min 5XX Error Rate5XX errors summed over one-minute intervals.
Disk UsageDisk space usage in gigabytes. The dashed line is the maximum available.

Collections

The Collections table shows the health of replicas across servers.

Garbage Collection

The following garbage-collection metrics are reported by Managed Solr:

MetricDefinition
ConcurrentMarkSweep CollectionTime & CollectionCountCollection time and count for the ConcurrentMarkSweep approach.
ParNew CollectionTime & CollectionCountCollection time and count for the ParNew approach.

Managed Solr provides graphs that report on throughput and latency.

MetricDefinition
Average Requests/second (ms)Reports average search requests per second.
Average Time/Request (ms)Reports average time for search request to fulfill the query.
5-minute Request RateReports search-request rate over a 5-minute interval.
15-minute Request RateReports search-request rate over 15-minute interval.
TimeoutsReports server timeouts, where the server did not respond to the request fast enough.
ErrorsNumbers of search-request errors.

Indexing Throughput and Latency

Managed Solr provides graphs that report on throughput and latency.

MetricDefinition
Average Requests/SecondReports average indexing requests per second.
Average Time/Request (ms)Reports average time for indexing request to fulfill the query.
TimeoutsReports server timeouts, where the server did not respond to the request fast enough.
ErrorsNumbers of indexing request errors.
Update Handler graphsReports various Update Handler metrics over a period of time, such as commits, autoCommits, deletesById, etc.

Caches

Managed Solr provides graphs that report on Caches.

MetricDefinition
QueryResultCacheThis cache stores ordered sets of document IDs — the top N results of a query ordered by some criterion. Multiple metrics report how the query result cache is being utilized to fulfill search requests. Metrics include hitratio, evictions, warmup time, etc.
FilterCacheThis cache stores unordered sets of document IDs that match the queries. Multiple metrics report how the filter cache is being utilized to fulfill search requests. Metrics include itratio, evictions, warmup time, etc.
DocumentCacheThe documentCache stores Lucene Document objects that have been fetched from disk. Multiple metrics report how the document cache is being utilized. Metrics include itratio, evictions, warmup time, etc.
FieldValueCacheThe fieldValueCache is similar to fieldCache, and supports multiple values per document. It is primarily used for faceting. Metrics include warmup time, size.

Service – Total Requests

Premium clients may view one more Pulse graph. The Service graph displays the number of requests received by the cluster’s Load Balancer over time. This can reveal cyclic patterns of high demand, as seen in the following graph:

The sum of requests in the chosen period of time is posted below the graph.

Availability Report

Premium clients get access to an Availability Report in the SearchStax dashboard. This is a graph and a summary that documents response time and downtime so they can be compared with the guarantees of the SLA.

Note that this feature is not a “Pulse” graph, but is accessed through the Availability Report screen of the Managed Solr dashboard.

Alerting

The Managed Solr alerting features are described on the Solr Service Alerting page.

Questions?

Do not hesitate to contact the SearchStax Support Desk.

Was this article helpful?
YesNo