Who will watch the watchers? Extended infrastructure observability for WSO2 API Manager

The Dynatrace WSO2 API Manager monitoring extension leverages the Dynatrace Davis AI to watch for unforeseen changes in underlying API responsiveness and automatically identify slowdowns in the performance of your API manager and point you to their root cause.

The technology race never stops. Those who lead it gain business advantage and market share. As in any race, sustainable advantage requires extensive preparation and attention to what’s under the hood. It’s cliched to say that cloud adoption has changed everything in this race, but a full understanding of the intricacies of cloud-native applications is still rare. Sure, cloud infrastructure requires comprehensive performance visibility, as Dynatrace provides, but the services that leverage cloud infrastructures also require close attention. Cloud-based application architectures commonly leverage microservices. Well-defined APIs are required for managing such microservices and tracking changes in their performance.

Extend infrastructure observability to WSO2 API Manager

Managing APIs in microservice environments, where component versions change dozens of times a day, is not easy. Add in the opportunities for distributing applications across multiple infrastructures—cloud, hybrid, or on-premises—and you can see how this challenge led to the rise of API management as a separate discipline, a discipline that’s critical to gaining a sustainable competitive advantage. In response to this trend, open source communities birthed new companies like WSO2 (of course, industry giants like Google, IBM, Software AG, and Tibco are also competing for a piece of the API management cake).

WSO2 API Manager became the leader in API management and is broadly used by Dynatrace customers to manage their APIs. WSO2 API Manager helps you to secure, govern, and analyze myriad incoming and outgoing API request types, but it offers little help in regards to its own operations.

Save hours of bug hunting with out-of-the-box WSO2 API Manager monitoring

The Dynatrace Software Intelligence Platform gives you a complete Infrastructure Monitoring solution for monitoring cloud platforms and virtual infrastructure, along with log monitoring and AIOps.

By leveraging the Dynatrace Davis AI causation engine to watch for unforeseen changes in underlying API responsiveness, Dynatrace automatically identifies slowdowns in the performance of your API manager and points you to their root cause. Specifically, Davis AI enables visibility into the following issues:

  • WSO2 AM services going dark
  • Soaring number of active connections
  • Stalled I/O of the underlying message bus
  • High latency or lack of responses

API manager monitoring from the application server perspective, which is what Dynatrace delivers with the WSO2 API Manager monitoring extension, can save you hours of bug hunting time. Read on to see how it works.

Find out why your API responses are slow

Imagine that it’s another busy afternoon at your office. You receive an alert message from Dynatrace (your infrastructure observability hub) letting you know that the average response latency of all deployed APIs has tripled. If something is not done soon, the whole ecosystem might be brought to a grinding halt.

But where does the fault lie? Is it the WSO2-AM gateway itself, a networking issue, a sudden increase in demand, or something else entirely? A quick answer to this question could make a world of difference.

Dynatrace integrates with WSO2 API Manager out of the box, allowing you to quickly see the root causes of such issues. In this case, Davis AI has identified that the root cause lies in an abnormal surge in API consumers—ordinarily, we would just trust our API manager and scale up API instances in response to such a discovery, but for this example, let’s double-check the diagnosis.

WSO2 API Manager latency chart in Dynatrace

Looking at the key metrics of the deployment does not reveal anything out of the ordinary. All WSO2 components seem to be working well.

WSO2 active connections chart in Dynatrace

Drilling down into the process metric details, the WSO2 API Manager (Overview) tab reveals the potential root cause of the issue: The count of active connections has been increasing quickly. This increase is clearly correlated with the increased response latencies.

No significant changes to message bus I/O times

Taking a look at a different WSO2 API Manager [WSO2 API Manager (Carbon)] allows you to confirm if the increasing number of clients is the source of the issue. As there are no significant changes to message bus I/O times, we may safely assume that the deployed APIs are not able to keep up with the demand or to take the appropriate corrective actions. This short journey through collected metrics prevented a serious issue and a long bug hunt.

Get a holistic overview of your WSO2 API Manager metrics

Dynatrace monitors 29 WSO2 API Manager–related metrics. These metrics give you a holistic overview of your WSO2 AM deployment health. The metrics are self-explanatory and intuitive out of the box, but here’s a short summary of the most important metrics:

  • WSO2 Message Bus throughput statistics—Should these be unusually high, investigate the persistence provider, as this often means issues with the disk. The metrics are: read/write response time (75th percentile) and read/write response rate in 15 minutes window.
  • Response latencies overview—Due to clearly marked “mediation” times for requests and responses, it’s quite easy to see if spikes in response times are the fault of WSO2 AM or APIs. The metrics are: average latency for total, average latency for backend, and response and request mediation. Predefined alerts are triggered when latency exceeds 30 seconds.
  • Active connections—These are separated into listener and sender categories, so you can quickly identify where the bulk of operations comes from. The metrics are: connection count per https/https listener and connection count per https/https sender.

For full technical details, metric reference, and prerequisites, see WSO2 API Manager in Dynatrace Help.

Why Dynatrace

Actionable application performance management requires breadth and depth. In fact, it’s a multidimensional discipline. Only a truly open, all-in-one platform with automated AI analytics at the core can analyze all applications, infrastructure, services, and enabling systems, along with their dependencies, in real time.

This is Dynatrace: the purpose-built Software Intelligence Platform, where API manager performance monitoring extends the dependency model of application transactions and infrastructure. Performance analytics are enriched with API manager insights into transaction performance and the root causes of detected performance degradation, using causation, not correlation, and delivering intelligent infrastructure observability with application context in mind.

Get Dynatrace today and avoid guesswork in facing the complexities of microservices.

What’s next?

This is the first public release of the Dynatrace WSO2 AM integration. In future releases, we’ll focus on responding to your feedback, so please share your thoughts and ideas for improvement with us at Dynatrace Answers.

Stay updated