You are on page 1of 5

Database Performance

Monitoring Explored
and Explained
Ed Tittel

CONTENTS IN THIS PAPER


Improving Database Performance  2 Database performance monitoring seeks to provide 24/7 database

Digging Into Database Health  3


health metrics and visibility into database queries and code, with


access to details from inside the database engine and from the host
Addressing Database Performance 3 operating system (OS) perspective. This provides numerous benefits,

Improving Database Efficiency  4


including the following:


• Past performance and behavior inform current observations
Measuring Code Deployment Impact 4
• Developers and DBAs iterate on code and query changes for contin-


Meet SolarWinds Database uous improvement and optimization
Performance Monitor 5
• Drive capacity and growth planning using statistical inferences and
projections

• Pinpoint and address performance and resource issues directly and


expeditiously

DATA BAS E P ER FO R M A N CE MO N I TO R I N G E X PLORE D A ND E XPLA INE D


1
Because data drives business success, anything organi- storage, and networking). These insights add consider-
zations can do to speed up troubleshooting for database able power to a company’s ability to create more efficient
problems is pure gold. Time saved in solving problems databases and improve existing database usage.
more quickly or delivering faster, more reliable access
Database performance monitoring provides a histori-
to data boosts productivity and profitability. Continuous
cal view of data so past performance and behavior can
database performance monitoring helps organizations
continuously inform current observations. This same
surmount troubleshooting hurdles and get more done.
historical view also supports trend analysis, growth and
This is more important than ever, as data is being cre- capacity planning, and ready identification of outlier
ated at a phenomenal rate—1.145 trillion MB per day. IDC values potentially indicating performance, reliability, or
estimates this daily number will grow to 463 exabytes by availability problems.
2025 (a multiplier of just over 404,000 in just four years,
for a CAGR of 2421%). Exponential growth like this intro- Improving Database Performance
duces massive complexity.
The deep and comprehensive visualization provided by a
It also changes how we must design, manage, and mon- top-shelf database performance monitoring tool forms
itor databases storing such data. Furthermore, proper a strong foundation for optimization, improvement, and
database management is an essential part of doing busi- problem-solving. An adaptive fault detection facility can
ness and helps ensure organizations can meet and exceed uncover previously unseen interruptions in the database
end-user expectations and in-house requirements. server or in database server availability. Given the right
Thus, it all comes down to maximizing and optimizing such tool, IT staff can pinpoint something like a MySQL
database performance. server stall fault, along with host ID on which the data-
base is running, its time of occurrence, duration, and so

Proper database management forth. Various details for such a fault appear in Figure 1.

is an essential part of doing In a similar fashion, DBAs, IT staff, and developers


can use a well-built database performance monitor-
business and helps ensure ing tool to take them deep into database performance
organizations can meet and details and data.
exceed end-user expectations
and in-house requirements.
Database performance monitoring also provides objective Fault Info
metrics permitting organizations to evaluate the perfor-
mance and resource impact of code changes to database Category: MySQL Server Stall Fault

applications and queries. Ultimately, this means more


Host: aps1
efficient and effective use of data and a more profitable
and productive organization. It also means extending the
Time: 11:05:17 AM Dec 12, 2020
purview of database performance management into the
cloud and across all database instances in use, no matter
Duration: 18
where they reside.

In particular, database performance monitoring zeroes in


Severity: 12.5

on database queries. It’s able to illuminate their comple-


tion times, identify factors contributing to latency, and
detail their resource consumption (including compute, Figure 1: Details for a MySQL server stall fault

DATA BAS E P ER FO R M A N CE MO N I TO R I N G E X PLORE D A ND E XPLA INE D


2
Digging Into Database Health about where and how to change server configurations.
Additionally, it can identify badly behaved queries
Database performance monitoring also provides 24/7
through its process list and other “top offender” or “top
real-time visibility into database health based on cur-
consumer” tracking tools.
rent measurements and metrics in light of prior perfor-
mance and best/average/worst-case comparisons. This
drives ongoing insights capable of spurring performance Addressing Database
improvements based on query tweaks, design optimiza- Performance
tions, and code changes.
To help organizations address database performance
Furthermore, a good database performance monitor
issues, a good database performance monitor lets users
helps simplify and streamline ongoing checkups with
track and plot metrics over time. It can illuminate trends
flexible, built-in tools (such as custom scripts for gather-
and changes within its health summaries for database
ing performance metrics) and expert recommendations.
instances and the systems hosting them. Figure 2, for
Better still, such software can uncover issues or ineffi-
example, shows the break-down for how MongoDB pro-
ciencies and offer remediations or improvements.
cessing time is split across multiple concurrent activities
Overall, a well-built database performance monitoring and clearly indicates that query processing is consuming
tool also helps organizations track and converge on key an overwhelming share of compute resources.
performance indicator (KPI) goals. These include targets
In addition, a well-built database performance monitor
for availability, latency, user experience metrics, and
can make recommendations based on best practices. This
improved mean time to detect (MTTD) and mean time to
helps organizations with key activities such as database
recover (MTTR) metrics.
and query optimization. Ongoing recommendations help
From an operations standpoint, a good database per- organizations keep improving their performance and
formance monitoring tool can make recommendations reliability constantly and continuously.

Figure 2: In this snapshot of current operations in MongoDB, query processing significantly outweighs other ongoing activities

DATA BAS E P ER FO R M A N CE MO N I TO R I N G E X PLORE D A ND E XPLA INE D


3
A good database performance monitor usually incor- and analyze trends and seasonal or periodic peaks and
porates tools to identify and investigate performance troughs in resource consumption and utilization. Thus,
outliers. By offering snapshots of current operations for a high-quality, high-capability database performance
a database platform like MongoDB, it can visually report monitor gives teams the insight and visibility into data
that query processing swamps other ongoing operations they need to successfully manage database performance.
by several orders of magnitude. Such displays point data-
base professionals directly where further investigation Measuring Code
and optimization are likely to lead to big payoffs.
Deployment Impact
In similar fashion, a good database performance monitor
In the same vein, developers, QA and DevOps teams, and
helps organizations measure resource capacity for data-
DBAs can use a well-built database performance moni-
base instances and the hosts on which they run. Then,
tor to see the impacts of new code on database workload
organizations can use those metrics to manage capacity
and query response prior to production deployment.
and match resource demand and supply. Thus, they can
Team members can compare query responses and overall
make sure applications and servers are properly provi-
database and application performance before and after
sioned to prevent or mitigate resource-based bottlenecks.
code changes to understand what’s helping and what’s
making things worse and iteratively run experiments to
Improving Database Efficiency improve outcomes.
DBAs, developers, and QA and DevOps staff can use a good
database performance monitor to visualize impact and
The deep and comprehensive
runtime for upcoming changes. Typical code tools might
include visual performance analysis, deep SQL parsing, visualization provided
and execution plan analytics. All such facilities combine by a top-shelf database
to help teams understand and optimize changes to exist-
performance monitoring tool
ing code and queries as well as new rollouts. Ultimately,
this leads DBAs and appli­cation teams (including DevOps forms a strong foundation for
staff) to keep databases healthy and proactively antic- optimization, improvement, and
ipate and forestall problems or issues. Clear objective
problem-solving.
metrics and observations help organizations achieve
better outcomes based on proven results.
DevOps and IT staff can use such a database performance
Careful attention to historical performance data in light monitor to compare top query performance over time
of current observations also helps database teams. They to drive toward continuous improvement. They can also
can more quickly identify events, problems, or incidents. find new queries with minimal effort (two clicks) so they
In the same vein, these teams can detect outliers or can easily zero in on items in need of analysis and pos-
anomalies before they escalate into noticeable delays or sible optimization. At any point, they can examine query
observable performance problems. details and performance, including sample results and
execution plans to determine what’s happening at run
Above all else, a well-built database performance mon-
time. Team members can spot events and trends this way
itor indefinitely maintains high-granularity historical
and steer development toward more optimal and effi-
data for purposes of comparison and low-granularity
cient implementations. The same holds true when code
data (which may be collected as often as every second)
and queries make the transition from test to production.
for more than a year. This lets users zoom into any point
in the historical record to gather observations driving
capacity planning, assure scalability as growth occurs,

DATA BAS E P ER FO R M A N CE MO N I TO R I N G E X PLORE D A ND E XPLA INE D


4
Meet SolarWinds Database An inside-out and outside-in perspective enables truly
comprehensive database performance monitoring, which
Performance Monitor
puts applications and queries into perspective within
SolarWinds Database Performance Monitor is a soft-
®
their host systems and the general database instances
ware as a service (SaaS)-based platform for monitoring inside which they run. Such metrics help distinguish the
traditional, open-source, and cloud-native databases. performance of applications and queries from other,
SolarWinds Database Performance Monitor offers a web- similar queries and applications, and it also helps dis-
based UI accessible on any internet-connected device, tinguish supporting services such as replication, backup,
and it works with all the well-known traditional, open- and load balancing.
source, and cloud-native databases, including NoSQL,
A lesser-known advantage of database performance
PostgreSQL, MongoDB, Redis, and more. Because it’s
monitoring is it promotes infrastructure cost savings.
cloud-based itself, this tool can monitor open-source
This is because it allows organizations to pinpoint and
databases running anywhere, whether they’re on-prem-
remedy specific causes for performance issues. They
ises, in the cloud, or in hybrid instances across multiple
need no longer “throw hardware at the problem” to solve
runtime environments.
their performance issues when more focused and precise
SolarWinds Database Performance Monitor makes remediation is readily available from the performance
both real-time and historical data available to its users, analysis tools and recommendations database perfor-
backed up by its own compute, storage, and networking mance monitoring puts at their fingertips.
resources. Thus, using this kind of database performance
monitoring imposes low overhead on databases in oper- GET STARTED TODAY
ation. It collects data the databases make available con- SolarWinds Database Performance Monitor is just
stantly and routinely, but it uses its own resources for the kind of database performance monitoring tool set
its built-in analysis, history, and reporting capabilities modern, forward-thinking organizations need. Please
without consuming database resources. visit the SolarWinds Database Performance Monitor
homepage, where you can request a demo or start a
free trial.
DevOps and IT staff can use a
well-built database performance
monitor to compare top query
performance over time to drive
toward continuous improvement.

This explains how SolarWinds Database Performance


Monitor makes current snapshots available alongside
historical data. It also explains how the tool provides
24/7 real-time database health monitoring, includ-
ing both internal database metrics and system metrics
from the operating systems inside which the database
instances run.

DATA BAS E P ER FO R M A N CE MO N I TO R I N G E X PLORE D A ND E XPLA INE D


5

You might also like