Invaluable, Trusted Performance Diagnostics
Updated December 04, 2019

Invaluable, Trusted Performance Diagnostics

Anonymous | TrustRadius Reviewer
Score 9 out of 10
Vetted Review
Verified User
Review Source

Overall Satisfaction with SolarWinds Database Performance Analyzer

DPA is an essential part of performance monitoring our hybrid Demand Management Solution for hosted customers, and select on-premise customers. Automation is essential considering the number of customers, the schema comprising approaching 3,000 tables and over 5,000 code artifacts. Daily summaries of the last 3 days of Top SQL resource consumers give us ongoing snapshots by which to spot trends and anomalies that DPA's functionality enables us to investigate to query and table level.
  • Automated Collection into a queryable repository - Hands-off, detailed collection assures us that the bulk of any performance issues are noted, so that their data can be presented on-demand in a readily consumable form. Drill-into features enable detailed analysis.
  • Performance Trend Analysis - Monthly top-consumers visualized in meaningful stacked-charts, with anomaly highlights and per-query, per-day detail enables hot-spot drill-down analysis.
  • Resource Utilization Warnings - Best-practice warning and critical measure applied to all facets, visualized in real-time are essential features that trigger troubleshooting.
  • Usability within the Trends and Current pages need improvement - Version 11 introduced a new Tuning feature, and changed the way one could quickly name a query and chose whether it would be of further interest, or is more infrastructure-oriented such that exclusion is warranted. Usability took a downward slide.
  • Resource Utilization has preset periods which, for longer durations, overly smooth charts. Custom periods are much needed via the UI, rather than repository-querying.
  • Storage analysis, similar to resource usage, needs custom periods and far greater ability to delineate between drives and files than v11 currently offers. (One rarely uses the Storage I/O tab)
  • One DBA can oversee performance issues for dozens of DB servers, serving to triage and disseminate using DPA email capabilities.
  • Development and Support personnel can see the same data, in the same visualizations, enabling skills-development on actual code and tables that would take too long to replicate as a purely teaching exercise.
  • Customer solution performance concerns can be directly investigated by time-slice. Additionally, using hands-on monitoring while customers replicate issues, our developers can see the artifacts in real-time that are contributing to the issues. Communication, visualization, identification, in real-time, obviates artificial replication efforts, saving time and money.
There are far too many to isolate a single, detailed, case. Suffice it to say that hour-long queries, down to many-per-second queries have been identified by DPA as top consumers. The mix of code-pinpointing and index usage (or the lack thereof) has enabled multiple hours of elapsed and CPU time to be cut for very high volume customers (row-counts over 10 Billion in a 1.2TB Database), such that very narrow SLA windows are conformed-to.
The Table Tuning feature is new in v11. It is sometimes used. Often its information is repetitive, making new revelations hard to spot - some "acknowledged" marker would help make this page more useful.
We use both DPA and Red Gate SQL Monitor (SQM) as third-party tools, and Query Store within SSMS. SQM's alerting and reporting capabilities exceed those of DPA in terms of being able to present single-purpose tiles for a server, where excessive smoothing doesn't hide some important detail. The SQM Estate page is particularly useful for space management and patching. DPA excels at query and table pinpointing - we see these tools as complementary. SQM is deployed for Production-only, whereas DPA is used for development through QA, UAT, and production.
Having used DPA for a multi-tenant, single DB server, SaaS, solution, and currently monitoring over 30 productions and QA DB servers, on a server-per-customer basis, SaaS solution, both are entirely appropriate DPA targets. Wherever lower spend due to optimized DB performance is considered, DPA is one of the best tools to meet that kind of objective.