AppDynamics Serverless APM for AWS Lambda

My usual use cases with Splunk AppDynamics  are primarily related to the Innova project, where we collect all assets and prepare a dashboard to display all assets and create reports for management about what assets are in place, their numbers, and their identifications.

We use Splunk AppDynamics  for application performance monitoring in on-premises environments because hybrid is not an option for us, as we utilize an OT system located on the client's premises, and we do not use a hybrid or any cloud environment.

We use Splunk AppDynamics primarily for maintenance and support, especially given the current threat landscape. We validate all software and hardware lists approved by the end user and monitor for any recently added assets to ensure compliance with the approved repository.

I rate Splunk AppDynamics' line-of-code level troubleshooting feature a perfect 10 out of 10 for diagnosing performance issues, as it is very good.

The codes used for Splunk are very good, understandable, and readable, which is why I give it a high rating.

I find Splunk AppDynamics' Digital Experience Monitoring feature very good because it offers end users complete insights into their network, hardware, and software assets, along with detailed reporting on asset installation, models, and operating systems.

The Secure Application feature in Splunk AppDynamics is valuable because it links to the firewall and reports any incidents or unexpected asset discoveries, thereby preventing unauthorized access or communication within the network.

I do not believe there are areas to be improved in Splunk AppDynamics rather than the current situation, but I believe that integrating artificial intelligence and linking the software with other APIs, such as AI or machine learning, would be great.

I wish to improve the AI and machine learning capabilities in the software, which are currently used for top management reporting instead of manual reporting from the system. I believe AI could provide more insights for annual or half-yearly reports and forecast future changes in the asset landscape.

I have been working with Splunk AppDynamics for approximately two years.

I give Splunk AppDynamics a 10 out of 10 for stability and reliability since it offers full insights into the software and hardware supply list.

Both stability and reliability are excellent in Splunk AppDynamics, with my assessment reflecting its consistent performance.

I have never experienced any outages with Splunk AppDynamics; it has never stopped working abruptly.

Scalability is very good in Splunk AppDynamics as it can accommodate a wide range of asset counts, but it depends on the license schema, which may range from 250 to 10,000 assets.

I normally communicate with the technical support of Splunk AppDynamics for any incidents or issues, but we have not opened any tickets yet. We also have 40 optional support hours committed in our contract.

Currently, I have not communicated with any technical support specialists from Splunk AppDynamics because we are still in the integration phase, but if we face issues during testing, we will use the support hours we have committed.

Before using Splunk AppDynamics, I used an outdated solution, an Excel sheet, to track all assets, software, and hardware, which was challenging compared to the resilience that Splunk AppDynamics' dynamic software provides.

My experience with the initial setup and deployment of Splunk AppDynamics is that it was very easy, with our team installing it in four control centers in less than eight days, which I find to be perfect.

We requested all technical manuals and third-party documentation from Splunk AppDynamics to deliver them to the end user as part of my commitment, ensuring they receive all necessary documentation and training materials.

The pricing and licensing of Splunk AppDynamics are managed by another team called the resource team, and they handle all tool and software lists to provide us with the necessary licenses.

I consider the pricing of Splunk AppDynamics to be mid-range compared to other software, as it operates on a subscription model, and the resource team finds the best prices, ensuring that Splunk AppDynamics is an acceptable option for the end user.

I evaluated other vendors, but we chose Splunk AppDynamics because the customer provided a list of approved software that we were required to procure and install without deviations.

I have experience with IT asset management solutions, specifically with Device42 , and also with asset discovery tools such as Splunk AppDynamics and other tools.

I was working in one of the projects in the Western region of Saudi Arabia at Neom city, where it is called the Innova project, and we are using Splunk AppDynamics for technical purposes there.

My role is as a delivery manager, and I am taking care of all software and hardware delivery at GE Vernova, where I have a team of integrators, network engineers, and database engineers. I ensure that everything has been delivered correctly to the end user and customer.

I am not in charge of getting into the technical details, but I follow up on the software names to ensure they are configured and in place, and I see that the automatic asset discovery feature, which automatically detects any additional hardware, is very interesting.

I think Splunk AppDynamics' data collection feature with agents or OpenTelemetry  is perfect because it is installed dynamically on all servers and nodes, and all agents respond quickly to the server or master station where Splunk AppDynamics is installed, providing great insights into hardware and software discovery.

Both stability and reliability are excellent in Splunk AppDynamics, with my assessment reflecting its consistent performance.

The alerting feature greatly impacts end-user experience since it is user-friendly and the end users are well-trained by GE, making them capable of understanding and operating the software without complications.

Splunk AppDynamics  allows us to understand the mean time to resolution and decrease it by easily understanding the dependency of the full application flow map. For root cause analysis and other hidden aspects, we can see how code quality performs. SQL queries can be easily evaluated for quality, and when code quality is not good, we can identify slowness in specific classes and methods. We can see which parts of methods and SQL queries are facing slowness issues. After that, we can develop and change code, modify database queries, and easily see in the product environment without needing to debug facilities. We can see in real time whether code changes are affecting the system positively or negatively. There are many different advantages, and we can separate proactive and reactive sides.

When we collect different method parameters with the essentials of monitoring APM  tools, we can easily combine business and operational development cycles in a single pane. For example, when development teams make process improvements to code to add new features to transactions, we can easily see how this feature affects customer experiences as performance metrics. If we can collect this kind of data, we can also easily combine business and operation metrics. For example, with a loan application from the customer side, such as a bank customer making a loan application over their mobile or internet banking application site, we can easily see how many successful transactions occur in real time from a business perspective, not just from the performance side. If we can collect these metrics, we can combine all performance and business metrics in a single pane, giving customer sites a very different and big picture view.

Splunk AppDynamics  works for Java, .NET, .NET Core, Node.js, and PHP applications. We also work with some customers using SecureApp features, and customer feedback has been really valuable for us. From my customers' experience with this feature, the feedback is really positive. In the software development lifecycle, penetration testing or security testing before getting a project into live production environment is a very important process. You have to conduct penetration testing before going live with your project. However, this kind of penetration testing is a bit reactive and offline because you only perform this penetration testing from a synthetic point of view, for example weekly, monthly, or quarterly. With Splunk AppDynamics SecureApp solution, we can easily see our application's vulnerabilities, attacks, and exploits in real time. We can also see any vulnerability, even zero-day attacks, easily after they occur. This is a really cool and differentiating feature, though it is a very new feature in the APM  market, almost two years old. Because of that, it is not well-known, but when we demonstrate it to customers in POC or demo sessions, most customers are impressed and want to try it in POC. After POC, some customers want to buy this feature while getting the APM solutions as well.

We can monitor what kind of vulnerabilities exist in the code and can easily show the business risk in the environment by making a business score, not relying on MITRE or CVE codes. The scoring also works from a business perspective. For example, if we have one vulnerability that may be medium severity, our internal scoring mechanism increases the business risk if the application touches databases or other inside applications. If the application is not communicating with other applications, databases, or other sources, the business risk may be lower than the other example because of the application's touching points. This is a really cool feature. We are not only reporting these vulnerabilities, but we are also blocking these attacks in real time. For example, when a Log4j2 vulnerability occurs on the system or any zero-day attacks happen, Splunk AppDynamics easily tags and understands this kind of attacks. If desired, it can easily block the application's attacks from the APM perspective. This is a really game-changer in my opinion.

I think one of the really strongest features of Splunk AppDynamics is the end-user experience monitoring. We have a really differentiated capability over our competitors. We can easily adapt our solution to the customer's application, internet banking solutions, or IoT devices all over the world. For example, when you get a new Volvo from any Volvo shop, that car has a built-in Splunk AppDynamics light agent to track their connected car applications. To give a specific example, Audi, Volvo, and four years ago BMW also use Splunk AppDynamics light agents to monitor IoT devices and connected car applications. Mercedes may have this kind of agreement as well. In summary, we can easily monitor mobile devices, including Android and iOS, browser-based applications, and also IoT devices. For example, in Turkey, I personally use IoT monitoring with my customers. We work with banks, and most customers monitor their ATM devices and POS devices via Splunk AppDynamics agents. I have personally implemented the IoT agent or Splunk AppDynamics agent into ATM devices and POS devices, as well as for some cinema companies' kiosk systems.

I can mention two different things. First, Splunk AppDynamics is mostly compared with the Dynatrace  solution because they are a really good solution. They offer on-premises options as well. I know Datadog  is another good solution, but it only works with SaaS solutions. New Relic  and Grafana  are also good solutions, but Splunk AppDynamics and Dynatrace  are the only on-premises options in the marketplace. Because of that, I want to compare with Dynatrace. Dynatrace has a OneAgent mechanism, while Splunk AppDynamics has a smart agent mechanism. The idea is quite similar, but when you use the Dynatrace OneAgent solution, because you are giving administrator and root rights, it is a bit easier but unsafe. For Splunk AppDynamics, you do not need to give the agent administrator or root privilege, but because of that, its capabilities are a bit limited. I cannot directly say this is a negative thing because it depends on the perspective. For example, if you really stick to security mechanisms, security teams can say that Dynatrace is easy to install and monitor, but from the security perspective, it is terrible and awful because you are giving full administrator and root privileges to Dynatrace. Splunk AppDynamics could improve their installation process, which would be an incredible thing on Splunk AppDynamics' side.

Second, most products, even Dynatrace, Splunk AppDynamics, and Datadog , are always saying they are making AIOps , root cause analysis, and anomaly detection features, but even Splunk AppDynamics, these kinds of features are not working fine because of the nature of the metrics. Most of the customers are not supplying the hygiene of metrics. If you do not supply or make your environment's metrics hygienic, you cannot give the AIOps  perspective to customers. The statement that these vendors can make root cause analysis automatically or have automatic detection features and capabilities cannot be truthfully said. To sum this up, this is not only a Splunk AppDynamics problem. From my personal perspective, this is all APM vendors' problem. The features that all these APM vendors need to improve are the AIOps features. These are really at the beginning of the AIOps era. Everyone is talking about AI, and it turns out to be a common hype in the technology market. We may see the real effect of this AIOps era maybe two or three years from today.

It is almost at the beginning of the story. When I started with Splunk AppDynamics, there was no acquisition between Splunk and AppDynamics, and AppDynamics was also its own company. This has been 14 years.

Implementation can be divided into two different parts. One part of the implementation process goes over the controller side, which can also be called the control plane side. One engineer is enough to install and prepare within two hours. However, the agent side is a bit more complicated because it depends on the customer's situation. If the customer has more than 100 or 1,000 different servers, mostly in production environments, we need to agree on when we deploy agents because we need to restart their applications. This creates some outages in their production environment. First, we need to agree on the timeline and project plan. It depends on the customer's decision. If we have a chance, at one time, we can also deploy more than 500 different agents at the same time, maybe within half a day, because we have really good playbooks and automation scripts working over Ansible , Chef , Puppet , or different automation tools that we can easily integrate. Implementation is easy, but the agent side depends on the customer's decision based on their project coverage or decision. It can take two days, or maybe two weeks or two months, depending on how big their environment is and how many licenses they get. For example, if they get more than 1,000 licenses for more than 1,000 hosts, it depends on their decision and project plan. It can take two months, one month, three weeks, or two weeks. It is a very variable thing.

I am working for my own business. Previously, I was in my former company, but I quit and built my own company. We are operating in the same area, and nothing has changed significantly in my life.

Over the last 14 years, we have made maybe more than 500 different installations, maybe much more than that. I do not know exactly, but in the last four years, we have prepared our own scripts and playbooks. It is really very simple to build a Splunk AppDynamics platform over the on-premises data center. Even if the customer wants to use the high availability option, if they have a limited environment or limited hardware resources, we can easily build all these components in one server. Because of that, it can be very simple when working only with one server. However, most customers in Turkey, including fintech-based, banking, government, and some really huge enterprises, need to use the high availability options, which means using more than six different separate servers. With our playbook and the solution's flexibility, Splunk AppDynamics is very flexible for this kind of model, and it takes no more than two, three, or four hours.

I would rate this solution an 8 out of 10.