Agent Data Is Just One of the Data Sets You Need for ITOA
Agent data is extremely valuable, but is only one of the data sets you need for complete IT visibility. Read how agent data fits with wire data, machine data, and synthetic data in an IT operations analytics (ITOA) practice.
Application Performance Management Comparison Criteria
AppDynamics, BMC, CA, Dynatrace, and New Relic all provide APM tools that can provide you with visibility into host-instrumented behavior. Based on the requirements of your organization, you should compare these APM tools in the following areas:
- Agent compatibility. Verify that the APM tools use software agents that are compatible with the hypervisor, application, O/S, and public cloud platforms that your organization relies on today or plans to use in the future.
- Deployment and licensing. The amount of visibility an APM tool deployment will provide depends on the number of hosts instrumented with agents, so you want to be able to scale the deployment easily and cost-effectively. A software-as-a-service (SaaS) model can help in this regard, with agents sending data back to a cloud-based analysis service run by the vendor.
- Programmability. You should be able to easily and programmatically configure the agents to collect specified data from transactions, including data that can only be gathered from the application code at runtime.
- Openness. Does the APM tool offer the ability to ingest third-party data for enriched insights? Conversely, you should look for an APM tool that allows agent data to be exported to an external data store so that you can correlate data sets and avoid vendor lock-in.
Wire Data + Agent Data
ExtraHop provides real-time analysis of wire data that complements agent data from APM tools. You would use the ExtraHop platform to gain a global, automatically discovered view of observed behavior for every tier of the application delivery chain—including components such as a WAN optimizer, router, firewall, switch, DNS server, LDAP server, load-balancer VIP, proxy server or memcached cluster, web servers, application servers, databases, and networked storage.
With this global view from ExtraHop across development, test, and production environments, organizations can compare actual performance against benchmarks for each tier to optimize performance and detect anomalies early. When you do see something worth investigating, wire data from ExtraHop will quickly guide to you to the right server to dig deeper into using agent data.
Here's a common example of how wire data and agent data work together: Looking at ExtraHop, the IT Operations team could see that a spike in web errors was correlated with a severe increase in latency for a particular database server, and then dig in to see the individual query statement that was performing especially slowly. With this information, the application support team could dig into the agent data gathered from that database server to see that a code change caused the slow down.
- Learn how ExtraHop and AppDynamics work together to provide transaction-tracing across the "opaque" network.