The network is designed to connect the organization’s users, partners, customers and visitors, but those connections are useless without software. While applications run on internal servers, end points and the cloud, the performance of the network in large measure defines the performance of the application, and this performance is what user experience and application experience (AX) is based on.
In this blog we’ll walk through network-based application performance monitoring (APM), why you should have it, and the best practices for making the most out of your network.
Ideally every app in your shop or accessible in the cloud should perform well and provide a positive user experience. But not all apps are created equal, and some, especially public-facing applications such as e-commerce, absolutely MUST perform at a high level.
Fortunately, application performance monitoring (APM) is there to provide deep insights into the user and application experience, and help IT comply with performance-related service level agreements (SLA).
At the same time, an APM gives IT information to identify bottlenecks and troubleshoot the applications. With this data and findings in hand, IT can finely tune applications to provide the best user experience that leads to satisfied customers and employees.
Why should you care about application performance monitoring? Short answer – you want (and maybe need) to know how fast your apps run, which ones are substandard, and how to get them up to par.
You especially want to know if customers and prospects are happy with the application speed or if they’re going over to your rivals because your customer-facing apps are too slow. Internally, you should know if employees using internal applications are productive and efficient or wasting time waiting for the software to respond.
If there is an application problem, IT needs to know the root cause of the performance issue so they can solve it.
Don’t Fall for Agent Shortcomings – Smarter Traffic Monitoring
Many APM systems rely on agents. On the plus side, agents can provide deep detail including specifics on the application source code. On the downside, agents are dependent on the operating system and application server. Due to the complexity, they can be expensive and time consuming to implement, and because they are installed across the network they can negatively impact the application performance.
On the other hand, traffic monitoring provides sufficient details to define and then tune the user experience and the application performance. Moreover, traffic monitoring is entirely independent of the operating system and applications. At the same time, agentless-traffic monitoring is passive and has no impact on infrastructure or applications – and can be deployed with a snap of IT’s fingers.
How Flowmon and Flow Data Work
Network flows are the secret sauce behind the five-star application experience recipe. Flowmon is a solution for network and security monitoring, and as the name indicates the solution is based on flow data which contains all the statistics about network traffic – essentially describing the communications in the network.
The first piece of the puzzle is Flowmon Probe, a standalone device which provides the network statistics in the form of flow data enriched with L7 data. These enriched flow data or network statistics are then exported to the Flowmon Collector which stores and visualizes the flow data helping IT analyze network traffic.
Flowmon Collectors are the heart of the solution, and like a server can be extended with software modules for different use cases and purposes. There are modules for NMI detection, network security, packet capturing, and of course the module for application performance monitoring this blog has been focusing on.
Deeper Inside Flowmon APM
Flowmon APM monitors the user experience and shows how long the user waited for a request to be responded to. APM measures these response times not only for the application but for databases and the network as well.
This solution is agentless – it simply sniffs the network traffic and monitors the responses. Using Flowmon APM, you can monitor all the modern HTTP and HTTPS applications. At the same time, IT can monitor transactions between the application and related database servers.
The APM module works on the application layer. It sniffs the traffic of the application, then does TCP assembly of the connection and measures the responses and other metrics.
How does Flowmon APM monitor transactions? Picture the user and the application server, and there between them is the Flowmon Probe monitoring the network traffic. When a user connects to the server, they send a request to which the application server responds and sends the response back.
APM measures the transport time of the request from user to application. More importantly, APM tracks the exact application delay — in other words,how long it took the application server to respond to the request. Conversely, APM measures the response from the application server to the user. When there are application delays, IT can view the details in the Flowmon APM console to see all response times for the applications and transport times for the network.
Transactions between the application and the database server are handled in a similiar manner, with APM measuring the SQL query transport time as well as the delay in the database server.
Meanwhile, once the database server responds, APM measures the transport time of response and shows this data and these delays in the console for IT to analyze.
Metrics and Charts
Ensuring a good application experience depends on measuring that application performance and comparing that to desired goals. This takes data that means something. APM data is made useful and actionable by being turned into metrics, charts and reports. APM metrics include:
Number of transactions
Server response time
Round trip time
Responses in maximum/minimum average percentile, and
Number of concurrent users.
All this can be synthesized into one killer metric – the APM index – a number between zero and 100 defining your application performance just the way your high school teacher graded your math - only with Flowmon, you can hope you a higher score.
Visualization and Troubleshooting
When application errors arise, APM provides IT and network pros error details such as the transaction error along with a timestamp, user identification, and information about the transaction.
To troubleshoot the application, IT can drill-down into the transaction and get general information about the transaction such as the response time, defined SLA, and associated username.
IT also gets information from the application layer: URL, network, cookie, status code, and other information from the HTTP header.
In short, Flowmon APM monitors the user experience using network traffic capturing and analysis and can distinguish between the network delay and application or database delay. APM monitors HTTP, HTTPS, and SQL databases.
Next, let’s look at how one enterprise harnessed AX.
Case Study – Health Insurance Company OZP Boosts Availability
OZP is the third largest employee insurance company in the Czech Republic. OZP offers customers access to all services and information through VITAKARTA ONLINE electronic portal and mVITAKARTA application for smartphones.
There are hundreds of internal application users and hundreds of thousands of VITAKARTA ONLINE application users. Given the criticality of the portal and the smartphone app, performance is a must. The company wanted to ensure both a high level of data network security and IT services availability.
Flowmon APM benefits for OZP include:
Integrated solution into current IT infrastructure surveillance system
Network traffic volume and structure reporting with 1 year history
Automatic detection and incident reporting
Real-time performance and response time monitoring
The VITAKARTA ONLINE self-service portal is where users get information and perform operations. The application runs in a datacenter and basically substitutes for branch offices, which is why it is so essential to ensure high availability and reliability of the application — this way it can be accessed by users with low response times and without application errors.
Before adopting Flowmon, OZP received user complaints about high response times of VITAKARTA ONLINE. With limited visibility it was difficult and time consuming for the IT department to identify problem sources and reasons for the applications’ response time deterioration.
It all changed with Flowmon APM. The APM solution now immediately informs IT about application performance deterioration and provides information about “where” and “when” applications had problems. The IT department knows about application performance deterioration even before users notice and allows them to quickly solve application problems, thus enhancing user experiences. Flowmon APM also offers reports for management and provides information about SLA compliance.
A second use is for internal applications that keep OZP in operation. The performance of internal applications influences work efficiency and the quality of services offered by OZP to its customers.
APM Does the Performance Trick
“VITAKARTA ONLINE application is used by tens of thousands of users of our insurance company. Application response time and user satisfaction is our top priority. Flowmon APM helped us identify bottlenecks, and by optimizating bottlenecks we achieved acceleration of response times by over ten percent. Flowmon APM provides us with evidence and reports about SLA compliance. The same applies for our internal application,” said Ing. Jan Devetter, Chief Information Officer.
Read our How OZP Identified Bottlenecks and Accelerated Response Time case study.