webMethods Insight

FAQs

Do I need to make application changes to start monitoring with webMethods Insight?
No. There’s no need to change your applications. Insight provides ready-to-use agents for monitoring a large number of application platforms. You can install these agents seamlessly on your application container. Without any code changes, you can start monitoring your applications right away.

How do I teach Insight the flow of my transactions and how to identify individual transactions?
Insight agents monitor the transactions flowing through the containers on which they are installed. Insight uses a patented Flow Mapping™ technology that automatically discovers services, consumers and maps interdependencies across numerous, mixed environments and protocols. Insight can automatically correlate messages that belong to a particular transaction as they move from one node to another.

Insight does this by adding transaction tracing information at the protocol layer. This correlation information is used to provide seamless auto-correlation capability across a service call chain that might involve multiple services hosted on different application containers. You don’t need to enable special debug options for each application nor add any transaction-specific information to a database for correlation purposes.

Will Insight change the content of my transaction messages?
No. Insight doesn’t need to change the transaction payload.

How will Insight impact the performance of my existing transactions?
Insight introduces less than five percent overhead on the services that are being monitored, even when volumes are high. Insight adds less than 50 microseconds of latency to each message. On-site tests show less than two percent CPU utilization. You can realize a message throughput of more than 3.5 million messages an hour on Insight managed nodes.

This is possible because Insight blends distributed policy evaluation with centralized command and control. In addition, because the Insight is never “in the flow of messages,” it will never detract from application performance.

How do I get visibility into a system that’s not managed by Insight?
Insight provides rich support for agents on the leading application platforms. We are constantly upgrading our breadth of platforms supported by agents. If a platform is not managed by Insight, you can still achieve visibility via:
1) Reconnecting flows—which give you visibility into traffic flowing through an un-instrumented node that’s sandwiched between two instrumented nodes.
2) One-hop visibility—for limited visibility into a next node (server) that’s not managed by Insight. Often this level of one-hop visibility is all you need.
3) A built-in Agent SDK kit—which you can use to build custom agents for platforms not managed by Insight.

What happens when the service call chain involves both synchronous and asynchronous calls? Is it possible to follow all the path and auto-correlate the transaction?
Yes. Insight agents don’t distinguish between synchronous and asynchronous calls. If all the servers (nodes) in the call chain are instrumented with Insight agents, then Insight can auto-correlate the transaction and provide deep visibility and monitoring for the transaction.

Can Insight monitor webMethods Broker?
Yes. Insight doesn’t need an agent to monitor the webmethods Broker. Simply use the reconnecting flows feature available in the latest version of Insight.

How is Insight different from traditional systems monitoring tools?
Insight and traditional systems monitoring tools solve two different sets of problems. Insight works at the transaction level and allows you to monitor each transaction flowing through your infrastructure. Insight provides SLA monitoring and root-cause isolation at the transaction level. In addition, with Insight, you gain cross-platform visibility for transactions that flow from one application container to another.

Traditional monitoring tools provide siloed visibility and monitoring at the system, network and application levels. Traditional systems monitoring tools offer little or no visibility into distributed business transactions. If a particular server or an application isn’t running, IT management tools will generate alerts but they can’t track the status of transactions that cross multiple systems.

How is Insight different from application profiling tools?
Insight and application profiling tools address different problems. Yet they are complementary when used together.

Insight helps IT map distributed business transactions and monitors their performance across systems. Application profiling tools, on the other hand, are used for managing and trouble-shooting traditional application server platforms, such as WebLogic and WebSphere. They focus on introspecting internals of an application server to tell you why a particular application isn’t performing.

Application profiling tools are too performance- and resource-intensive to run all the time. In contrast, Insight can be kept “always-on” in the production environment to identify problematic systems/applications.

We suggest using Insight to pinpoint the application layer, service name and transaction payload that caused a transaction exception and then using application profiling tools to look inside the application/service to locate the exact internal problem that’s causing the error.

Is Insight integrated with CentraSite?
Yes. Insight provides out-of-the-box integration with CentraSite ActiveSOA. Insight:
1) Identifies rogue services and publish them to CentraSite
2) Discovers run-time relationships and publishes them to CentraSite
3) Publishes run-time service metrics to CentraSite
4) Publishes Insight-discovered consumers to CentraSite
5) Publishes SLA violations to CentraSite