Kong Enterprise 0.29 is Released With Vitals Monitoring and Analytics Features
In October we introduced Kong Enterprise Edition, the microservice API platform for large organizations. Now, just about a month later, we’re delivering new features specifically for enterprise customers. Release highlights include:
- New Vitals Analytics Interface
- New Forward Proxy Plugin
- Improved Proxy Caching Plugin
Enterprise customers are encouraged to visit the Kong Enterprise Customer Success portal (subscription required) for the full list of features, enhancements, fixes and system requirements.
New: Kong Vitals Monitoring and Analytics
Kong Enterprise 0.29 introduces Vitals, Kong Enterprise’s new monitoring and analytics interface. Vitals lets you keep track of the activity volume of your API application and Kong instances, all from from a web browser (and soon via the Kong Admin API too).
Importantly, and like all Kong features, Vitals is designed to have the smallest possible impact on API performance. As you may know, in most scenarios, Kong adds <1ms latency to API requests.
Vitals has a visual interface in the Kong Admin GUI, so let’s give it a look:
This single screen provides a powerful view of API gateway activity over time. You can choose a desired time range and reporting interval, and Vitals displays current and historical results. Let’s drill into the three graphs:
- L2 Cache Hit/Miss. This graph shows accesses to Kong’s datastore cache. The blue line shows hits; the red line shows cache misses, which result in database fetches. There was an interesting moment at 12:25:03 pm when both hits and misses are relatively high. In this case the absolute number of cache hits and misses calls attention to this moment.
- L2 Cache Hit Percentage. The green line shows that Kong is utilizing cache most of the time. The period of time we focused on from the Hit/Miss chart shows as a 69.3% hit percentage. About 5 seconds later, the cache hit percentage chart goes to 0%, presenting a different but equally interesting moment. At that time the total number of requests are low, but nearly all requests require accessing the data store. The Cache Hit Percentage Chart surfaces this anomaly for investigation. Shortly thereafter the cache hit percentage climbs back to the near 100% level.
- Proxy Latency Min/Max. The yellow line shows the maximum latency introduced by the gateway; the purple line shows the minimum latency introduced at the gateway. Kong is designed to utilize cache whenever possible. This chart provides insights into how expensive certain cache hits are to overall performance. As expected, latency is zero except when new actions require loading information from the data store into cache.
Vitals data is is available for visual analysis in a web browser today, and will be writable to log files and 3rd-party monitoring platforms soon.
New: Forward Proxy Plugin
The Forward Proxy Plugin allows Kong Enterprise to connect to intermediary transparent HTTP proxies (instead of directly to a specific upstream_url in the API definition) when forwarding requests upstream. This is useful in environments where Kong sits in an organization’s internal network, the upstream API is available via the public internet, and the organization proxies all outbound traffic through a forward proxy server.
Improved: Proxy Caching Plugin
The Proxy Caching Plugin for Kong Enterprise makes it fast and easy to configure caching of responses and serving of those cached responses to matching requests. This release adds support for cache-control directives, easier-to-use Admin API endpoints to list and clear proxy cache data and more control over cache purging in multi-node Kong clusters.
And Lots More
Periodic Kong Enterprise releases consolidate new features, improvements, and fixes across all product areas to make API operations more feature-rich, reliable and performant. Kong Enterprise subscribers are strongly encouraged to review all changes and upgrade to this valuable new release.