Product

AppDynamics Enhances Pivotal Cloud Foundry Performance Monitoring with New Infrastructure View

By | | 4 min read


Summary
We’ve significantly expanded our Pivotal Cloud Foundry (PCF) performance monitoring capabilities, adding an intuitive dashboard, new metrics, and advanced capabilities to deliver powerful insights into your PCF applications.

AppDynamics’ strong commitment to Pivotal Cloud Foundry’s opinionated platform philosophy makes it easy to add our industry-leading APM to your PCF applications. Our new Service Broker Tile, available on PivNet, is the latest example of our commitment to the most advanced Pivotal architecture.

The tile was completely rewritten to support all the latest Pivotal Service Broker APIs, and now adds support for Cloud Foundry’s Loggregator System for detailed metrics on the health of the underlying infrastructure.

Full PCF Support

Many AppDynamics customers are moving to PCF 2.x to take advantage of the new platform.

We’ve refreshed our implementation of the tile—not only to support PCF versions 2.0 and 2.1, but also to make the tile lightweight, faster to install, and easier to use.

Using our open extensibility framework, we’ve developed a custom nozzle to pull all the important KPI metrics from the Logregator System—and we’re packaging this up as a BOSH Release for tight integration with PCF.

Our Dashboard and PCF Metrics

Administrators need an easy way to visualize—and receive alerts on—the state of their PCF environments. Unfortunately, it’s often difficult to get a clear view of these key metrics. The PCF KPI Dashboard provides that clear view.

Pivotal has very good documentation on the metrics they provide. These metrics are split between performance indicators and capacity scaling indicators. Our new dashboard encapsulates all of these metrics, providing a comprehensive view in a single pane of glass.

At the lowest level, there are health- and resource-related metrics for each of the services running. At a higher level, the KPIs focus on monitoring capacity.

The KPIs we show in our dashboard are from Pivotal’s recommended list. We ingest and publish these KPIs to the AppDynamics Controller.

PCF KPI Dashboard: An Overview

Our dashboard works with Pivotal’s comprehensive set of recommended metrics. Here are some examples:

  • Diego cell metrics: remaining memory available (cell memory chunk and overall remaining memory), remaining disk available, cell rep time to sync, unhealthy cells.

  • GoRouter: router file descriptors, router exhausted connections, router throughput.

  • System (BOSH): VM health, VM memory used, VM disk used.

Here’s a sample PCF KPI Dashboard:

  • The top-left section shows Key Capacity Scaling Indicators, including Diego Cell memory capacity, disk capacity, container capacity, and so on.

  • Top-center to right: Key Capacity Scaling Indicators are plugged into widgets, giving administrators an easy-to-visualize, graphical representation of the state of their PCF environment.

  • The middle section shows a series of KPIs, plus a widget for the current total number of routes registered with the Gorouter.

  • Near the bottom, we see the services that make up the PCF environment, including core metrics for CPU, memory, disk, etc.

For more details on PCF KPI Dashboard installation and configuration, go to https://docs.pivotal.io/partners/appdynamics/index.html  

The Loggregator System

The Loggregator gathers and streams logs and metrics from user apps in a Cloud Foundry (CF) deployment, as well as metrics from CF components.

AppDynamics was one of the first vendors to come out with a Service Broker tile to allow easy configuration of APM agents with Cloud Foundry deployments. Today, our tile is one of the most downloaded extensions from the Pivotal Network. Now we’ve completely refreshed the implementation and replaced the older JMX Bridge, a more limited way of extracting metrics from a PCF system.

Our new tile uses the Loggregator, showing our commitment to the latest PCF platform enhancements.

The Loggregator offers many advantages. Administrators, for example, can access the Loggregator Firehose, the combined stream of logs from all apps, as well as the metrics data from CF components.

If you’re not familiar with the Firehose, it’s a WebSocket endpoint that streams all event data coming from a CF deployment. This data stream includes logs, HTTP events and container metrics from all applications, and metrics from CF system components.

Custom Nozzle

Operators can deploy nozzles to the Firehose. A nozzle is a component that monitors the Firehose for specified events and metrics, and streams this data to external services.

We’ve built our own nozzle—a standard AppDynamics extension packaged as part of the Service Broker Tile—enabling us to pull metrics from the Firehose. This nozzle gives us access to all of the main metrics that administrators need to monitor a PCF system.

Our new Service Broker Tile, with its powerful dashboard, faster performance, and Loggregator support, is the latest example of AppDynamics’ strong commitment to the Pivotal Cloud Foundry platform. Download it today!

Mark Prichard, co-author, is Senior Director of Product Management at AppDynamics.