By If you’re hosting your applications on the Pivotal Cloud Foundry (PCF) platform, you’re already delivering better, faster software than ever before. But what about monitoring the performance of the PCF platform itself? We’re excited to introduce the New Relic Firehose nozzle, which provides access to critical PCF platform performance metrics.
The PCF Firehose “is a websocket endpoint that streams all the event data coming from a Cloud Foundry deployment.” If you’re an operator or administrator of any PCF components, you’re probably highly interested in event data like error logs, HTTP events, and metrics about containers. Like the New Relic Service Broker for PCF—which you’re hopefully using to monitor all your PCF applications—the New Relic Firehose nozzle formats all this data for ingestion into New Relic Insights so you can access it quickly and efficiently.
Dashboards, dashboards, dashboards!
Here’s a look at a sample overview dashboard from a running nozzle:
This dashboard shows the event load of all PCF domains, and provides further granularity by breaking down event types for each domain. New Relic customers can contact their representative for information on how to obtain pre-built dashboards for the New Relic Firehose nozzle. [click to enlarge]
- Event Types
- Jobs
- Origins
- Deployments
- HTTP transactions, including HTTP Status Codes, HTTP Request Types, and HTTP Peer Types
You can even drill into your container deployments and focus on a group of containers or even just inspect a single container.
You can inspect all container metrics or drill down into specific containers. [click to enlarge]
Included event types for the New Relic Firehose nozzle
The New Relic nozzle ingests all PCF Firehose event types into one Insights event type called PcfFirehoseEvent. You can view these events as the following metrics:
- Value Metrics: Provides data from any PCF component (for example, Cloud Controller, Diego, Doppler Server, MetronAgent).
- Counter Events: Tracks events for failed and started long-running processes (LRPs), failed and started tasks, number of HTTP messages received and sent, and number of bytes sent.
- Container Metrics: Tracks CPU usage, storage and memory usage, and memory and disk quota for each container running on PCF.
- HTTP StartStop Messages: Provides details including start and stop time of requests, IP address of message, and length of the message’s content.
- Log Messages: Records detected events, actions, errors, or any other user-generated messages.
This widget shows the number of events ingested over a period of 60 minutes.
(Tip: The New Relic Nozzle for PCF documentation has some sample Insights queries to help get you started.)
You can generate events from any of the following Diego metrics, which have the following origin names:
For details about these metrics, refer to the Diego metrics reference.
The following is a sample list of available PCF job types:
(Note: Depending on your PCF deployments, you may have more jobs available.)
It’s time to get started monitoring your full PCF deployment
You can now put your visibility into your usage of the PCF platform side by side with information about your application performance, for a complete picture for everything PCF!
The New Relic Firehose nozzle is available from the Pivotal network or from GitHub, and either link includes everything you need to know about installation and configuration. Currently, the nozzle is compatible with PCF versions 1.8 through 1.12.
To install it, you can either fill out the app’s manifest and deploy it as a regular application, or you can install the tile version in PCF Ops Manager so others in your organization can have easy access to all the data you’re about to gather. If you have questions or need assistance, contact your New Relic account executive or technical account manager.
Happy nozzling!
By