Welcome!

Recurring Revenue Authors: Yeshim Deniz, Liz McMillan, Elizabeth White, Zakia Bouachraoui, Pat Romanski

Related Topics: Recurring Revenue, Java IoT, Machine Learning

Recurring Revenue: Article

How to Monitor Oracle Database Performance

What metrics are we interested in?

An Oracle database provides several v$ views to query information about the database instance, including statistical information that can be used for monitoring and problem analysis purposes. Rene Nyffenegger wrote a nice Summary on Oracle’s v$ views that gives an overview of all available views.

The following illustration shows a dashboard with key metrics that we pull out from an Oracle database when doing performance management with our clients:

dynaTrace Dashboard showing key performance metrics queried from an Oracle Database via v$ tables

dynaTrace Dashboard showing key performance metrics queried from an Oracle Database via v$ tables

In this article I provide a quick overview on what these metrics tell us and how to get them. If you run dynaTrace in your environment you can also download the [DL:Oracle Monitor Plugin] from our Community Portal .

I also appreciate feedback from all Oracle experts out there. I am sure there are many more metrics that are interesting and important for performance analysis, so please add those here.

What metrics are we interested in?
Every time you start up an Oracle instance the system allocates memory for its System Global Area (SGA) (Read more on Oracle SGA Concepts). A very interesting area involves internal data buffers. These buffers hold data in memory and are first searched when a request comes in before fetching data from disk. Buffering obviously saves on I/O and speeds up database requests. There is a great article that explains the different buffer pools, how they can be configured and it also gives recommendations on sizing: Using the dynamic SGA Features of Oracle 9i.

Buffer, Execution and I/O metrics
There are several great blogs that explain metrics such as Buffer Cache Hit Ratio, In Memory Sort Ratio,Parse to Execute Ratio or talk about I/O metrics and how to tune your database based on metrics retrieved from v$buffer_pool_statistics.

The metrics that we query therefore are Buffer Cache Hit Ratio, In Memory Sort Ratio, Parse to Execute Ratio, SQL Area Get Ratio, Buffer Busy Wait, Free Buffer Waits, Write Complete Wait, Consistent Gets, DB Block Gets and Physical Reads.

Connection and User Count
The number of connections and user sessions are key system and performance indicators. The number of connections is usually configurable via connection pool settings on the application server. The more connections the more resources you need on the database - on the other hand you can serve more concurrent users that request data from the database. The blog v$license view tipsgives a good overview of values exposed by this view.

The metrics that we query from this view are Maximum Concurrent User Sessions, Current Concurrent User Sessions, Highest Concurrent User Sessions and Maximum Named Users

Connection Time
Additionally to the metrics that we query from the system tables, we also monitor how long it takes to actually establish a physical connection to the database. Our monitors use the Oracle Database driver and measure the time it takes to get a connection. This metric gives us a good indication on how good the database can deal with new incoming connection requests.

How to query them
Querying these values is pretty straight forward. [Our monitor] is implemented in Java. We load the Oracle JDBC Driver and establish a connection to the Oracle Database Instance when we initially launch the monitor. Check out the following blog with an example on How To connect to Oracle with JDBC. Then we simply execute the SQL statements to retrieve the actual measures on a scheduled interval (e.g.: every 30 seconds).

Most of the statements we execute are explained in the blogs I linked to earlier in this article, e.g.: Buffer Cache Hit Ratio, In Memory Sort Ratio,Parse to Execute Ratio

How to run a monitor and how to read the values
We run our monitors on the dynaTrace APM Platform . This means that these monitors are executed on a scheduled interval, e.g.: every 30 seconds. A monitor itself is a Java OSGI plugin with an execute method that queries and returns the metrics from the database views. The monitored values can then be displayed in a dashboard (as seen in the screenshot in the top) or can trigger alerts, e.g.: notify my admins in case Buffer Cache Ratio drops under a defined threshold.

If you implement your own monitor you need to figure out how granular you need the data and you have to figure out how to display and process the individual measure points.

How to read these values?
The most important thing to understand is to not only look at individual metrics but really look at your system - and with that I mean looking at measures from all application components that are involved to process individual transactions. A low Buffer Cache Hit Ratio doesn't mean that your DBA's should run off right away trying to increase this value by tweaking the db_cache_size. The business function of the application largely dictates the way it interacts with its backend datastore.  For example, if the application is more Decision Support oriented, effective use of the Buffer Cache may be impractical or impossible due to the large data sets that are being analyzed.  On the other hand, if the application is OLTP oriented, the Buffer Cache should be highly utilized and the behaviour of the Hit Ratio with respect to transaction mixes, peak concurrencies, etc ... becomes a critical measure to understand.  Nonetheless, even in these cases, you need to look at overall Transaction Response Times and see if you actually have a problem that affects the end-user.

If that is the case you need to figure out whether the time is spent in the Application Layer (bad performing code), the Network (too much data being sent between components) or the Database (inefficient settings or non optimized indices, ...). If it turns out that the problem is mainly caused by the Database Layer you can check whether your connection pools are exhausted. If that is the case you actually want to talk with the application developers and see whether they can optimize the usage of connections. If you have too much traffic on the network because too much data is requested from the database you want to look into optimizing the application code to really only query the data that is needed. If it then turns out that the problem is on the database you have to observe the metrics discussed in this article. There are ways to tweak database settings to improve buffer usage. I am not the expert on Oracle but if you follow the links that I've posted throughout you will find many good recommendations on what to do in various scenarios.

The database is most often not the root cause of transactions that perform poorly from and end-user’s perspective. Too often it is the application that is e.g.: keeping connections open for too long or is querying more data than needed. I recommend reading Top 10 Performance Problems taken from Zappos, Monster, Thomson and Co. Problem Patterns #1, #4 and #7 often lead people to point to the DBA's first instead of identifying these problems in the application layer. That being said - there are of course scenarios where it really is the database that slows down transaction response times. And in these cases it is important to have enough information available to analyze the root cause of the problem. Correlating the database metrics with other metrics from the underlying operating system, network, application server, web server, ... will make it easier to find a solution to your performance problems.

What's your strategy on database monitoring?
This is one example on monitoring your database. I am interested in getting your feedback on which tools and approaches you use to monitor your Oracle, SQL-Server, MySql, ... databases. Let us know (via comments - login required) which measures you monitor and you consider important. Thanks

Related reading:

  1. Lessons Learned: How we Monitor our Community Portal from the Cloud Our dynaTrace Community Portal is our gateway to our users....
  2. Apache Web Server Status Monitoring with a dynaTrace Plugin provided by MCG Systems The extensible plugin architecture of dynaTrace opens many doors for...
  3. Automated Performance Analysis: What’s going on in my ASP.NET or ASP.NET MVC Application? I’ve spent some time in the last weeks playing with different...

More Stories By Andreas Grabner

Andreas Grabner has been helping companies improve their application performance for 15+ years. He is a regular contributor within Web Performance and DevOps communities and a prolific speaker at user groups and conferences around the world. Reach him at @grabnerandi

IoT & Smart Cities Stories
After years of investments and acquisitions, CloudBlue was created with the goal of building the world's only hyperscale digital platform with an increasingly infinite ecosystem and proven go-to-market services. The result? An unmatched platform that helps customers streamline cloud operations, save time and money, and revolutionize their businesses overnight. Today, the platform operates in more than 45 countries and powers more than 200 of the world's largest cloud marketplaces, managing mo...
At CloudEXPO Silicon Valley, June 24-26, 2019, Digital Transformation (DX) is a major focus with expanded DevOpsSUMMIT and FinTechEXPO programs within the DXWorldEXPO agenda. Successful transformation requires a laser focus on being data-driven and on using all the tools available that enable transformation if they plan to survive over the long term. A total of 88% of Fortune 500 companies from a generation ago are now out of business. Only 12% still survive. Similar percentages are found throug...
When Enterprises started adopting Hadoop-based Big Data environments over the last ten years, they were mainly on-premise deployments. Organizations would spin up and manage large Hadoop clusters, where they would funnel exabytes or petabytes of unstructured data.However, over the last few years the economics of maintaining this enormous infrastructure compared with the elastic scalability of viable cloud options has changed this equation. The growth of cloud storage, cloud-managed big data e...
Your applications have evolved, your computing needs are changing, and your servers have become more and more dense. But your data center hasn't changed so you can't get the benefits of cheaper, better, smaller, faster... until now. Colovore is Silicon Valley's premier provider of high-density colocation solutions that are a perfect fit for companies operating modern, high-performance hardware. No other Bay Area colo provider can match our density, operating efficiency, and ease of scalability.
At CloudEXPO Silicon Valley, June 24-26, 2019, Digital Transformation (DX) is a major focus with expanded DevOpsSUMMIT and FinTechEXPO programs within the DXWorldEXPO agenda. Successful transformation requires a laser focus on being data-driven and on using all the tools available that enable transformation if they plan to survive over the long term. A total of 88% of Fortune 500 companies from a generation ago are now out of business. Only 12% still survive. Similar percentages are found throug...
As you know, enterprise IT conversation over the past year have often centered upon the open-source Kubernetes container orchestration system. In fact, Kubernetes has emerged as the key technology -- and even primary platform -- of cloud migrations for a wide variety of organizations. Kubernetes is critical to forward-looking enterprises that continue to push their IT infrastructures toward maximum functionality, scalability, and flexibility. As they do so, IT professionals are also embr...
ScaleMP is the leader in virtualization for in-memory high-end computing, providing higher performance and lower total cost of ownership as compared with traditional shared-memory systems. The company's innovative Versatile SMP (vSMP) architecture aggregates multiple x86 systems into a single virtual x86 system, delivering an industry-standard, high-end shared-memory computer. Using software to replace custom hardware and components, ScaleMP offers a new, revolutionary computing paradigm. vSMP F...
At CloudEXPO Silicon Valley, June 24-26, 2019, Digital Transformation (DX) is a major focus with expanded DevOpsSUMMIT and FinTechEXPO programs within the DXWorldEXPO agenda. Successful transformation requires a laser focus on being data-driven and on using all the tools available that enable transformation if they plan to survive over the long term. A total of 88% of Fortune 500 companies from a generation ago are now out of business. Only 12% still survive. Similar percentages are found throug...
At CloudEXPO Silicon Valley, June 24-26, 2019, Digital Transformation (DX) is a major focus with expanded DevOpsSUMMIT and FinTechEXPO programs within the DXWorldEXPO agenda. Successful transformation requires a laser focus on being data-driven and on using all the tools available that enable transformation if they plan to survive over the long term. A total of 88% of Fortune 500 companies from a generation ago are now out of business. Only 12% still survive. Similar percentages are found throug...
As you know, enterprise IT conversation over the past year have often centered upon the open-source Kubernetes container orchestration system. In fact, Kubernetes has emerged as the key technology -- and even primary platform -- of cloud migrations for a wide variety of organizations. Kubernetes is critical to forward-looking enterprises that continue to push their IT infrastructures toward maximum functionality, scalability, and flexibility. As they do so, IT professionals are also embr...