Application landscapes are complex and the complexity is increasing even further because we are not only connecting more and more apps, but also more and more different architectures. We integrate custom software and PeopleSoft with Microsoft platforms such as Microsoft 365 and Power Platform. This is how we mix SaaS, IaaS, PaaS and on-prem software together. It's complex, but it works. And, because we use all kinds of smart tools from all kinds of forward-thinking tech parties, we can innovate quickly and build great customer experiences.
And then your customers start calling. Because 'the app is slow'.
“Hope someone else solves it”
And that's where the trouble starts. Because your customer may imagine themselves to be a user of 'the app', we know that app is in reality a mixed landscape of point solutions, APIs and cloud platforms that is not only complex but also changing rapidly. And all those links in the chain can be the cause of a performance problem. Originally, there were always two ways to tackle such a complicated performance problem: a positive-minded manager with a sense of responsibility picked up the phone and called all departments and suppliers until the problem was resolved, and then came up with a solution. Another approach, which we unfortunately still see a lot, is “finding that it's not our fault and hoping someone else will fix it”.
These methods will no longer be acceptable in 2024. What you need is a mechanism to analyse the performance data of your entire chain very quickly. That's why we use Elastic.
What the Elastic stack is and how we use it
The Elastic Stack (or: ELK Stack) is a collection of open source tools for real-time search, analysis, and visualization of large amounts of data. The name ELK comes from the three main components: Elasticsearch, Logstash, and Kibana.
Elasticsearch itself is the search and analytics engine that forms the backbone of the stack. In Elastic, you can store and analyze huge data sets and perform complex queries quickly. Logstash and Elastic Agent give you a data processing pipeline that collects, transforms, aggregates and forwards data from various sources to Elasticsearch. Kibana is the dashboard part of the stack. That's where we build data visualizations that help administrators quickly see where a performance problem is or is imminent.
Elastic is not specifically intended, but it is very suitable for monitoring the performance of complex application chains. For full observability, we combine different data sources:
- Application and infrastructure log files. Here we find errors, warnings and other important events
- Performance metrics. Variations in CPU usage, memory usage, network traffic, and disk activity can all be indicators of performance issues.
- Application performance. Transactions, slow queries, and error messages give us insight into the user's perspective so we see issues before they call
- Synthetic Monitoring. Continuously simulating the most important customer journeys yourself provides a lot of information about how an application performs for the user. To do this, Elastic Synthetic Monitoring simulates the end user's browser.
This is how the Elastic stack is designed to be a powerful platform for observability, with which we can not only detect and solve performance problems, but also proactively prevent them.
“What does this data actually mean?”
Installing and setting up the entire Elastic stack, setting up the data pipelines, dashboards and triggers and then maintaining the environment may be specialist work, but we now have a lot of experience with that and it usually does not present us with major challenges.
But the technical questions aren't the most important questions at all either. Because we can set up a technically perfect environment for you, ultimately you have to decide what you want to monitor and for what purpose. And then you have to determine which actions should follow if the chosen metrics exceed a limit value.
You can collect a lot of data and put it on a dashboard, but it's no use without the answer to the question: “what does this actually mean?” In other words: observability is not an end in itself; there must always be a business goal behind it.
Observability with Elastic in practice
To show you how that can work, we'll give you a few examples from our practice.
- Better end customer experience
One of our customers had no insight into how their own application performed with consumers and was therefore only able to respond to problems when people started complaining. In digital services, where the application mission critical is, you can't afford that risk. By implementing observability with Elastic, this company was able to see for itself when things were slow on the user side and intervene proactively.
- More efficient management
Another customer we worked with had an internal reason for investing in better observability. This is because administrators spent a lot of time looking for the cause of problems in their network of technical components. That cause was often not found either, so it was completely unclear why the environment was sometimes slow. This was a source of constant worry and stress for the entire IT departments. With Elastic, they now find the cause much faster and can also prevent problems before things go wrong.
- Compliance
We also regularly help customers who don't so much have performance problems, but who need to be able to demonstrate through internal or external rules that they have control over what happens in their environments. They sometimes need to be able to show in real time what dependencies, bottlenecks and risks are in their application chain. Here, Elastic therefore serves more as a reporting tool for demonstrating compliance.
- Predicting the effect of change
With sufficient data about your environment, you can predict what a change will do to your performance and where upgrades or adjustments may be needed. Sometimes this involves a migration or about supporting a company's growth goals (“can we handle those 10,000 extra users?”) , but we can also provide insight into the impact of large amounts of IoT data released by innovative projects.
- Cost optimization
Performance and costs are inseparable. So many of our customers want to understand how far they can reduce costs by “scaling back” their systems without bothering the end user. This is a subtle balancing act that you can only do well with real-time data and really good dashboards. So it's a good application of the Elastic stack.
What is your performance challenge?
These are 5 challenges that we have helped customers with. But your performance challenge is probably just a little bit different. Because we always want to learn, we'd love to hear what performance questions you have. Then we can think about how Elastic could answer that.
Get in touch with:
Cees Schrijen, Account Manager Enterprise IT
c.schrijen@blisdigital.com | 06-17503107
Learn more about our Performance and Monitoring Services
.jpg)
