Contributions from Dinesha Sharma and Alka Gupta
Pivotal Container Service (PKS) comes with out of the box integration with Wavefront for monitoring and analytics of your cloud native Apps and infrastructuire, inclusive of the kubernets platform, containers, applications and services. In this blog, we call out some specific use cases that developers and platform reliability engineers (PRE) may find valuable for monitoring and analyzing the PKS platform along with its workloads and microservices using Wavefront.
1. Query driven analytics
Narrow down search using time series query language with respect to parameter like CPU, memory usage of a PKS namespace, node, pod etc.
For example:
To view the metrics for pod usage of a cluster named “VMwareWavefront”, create a time series query as detailed below:
Create a new dashboard in Wavefront portal as shown in figure 1.
Navigate to: Wavefront UI -> Integrations- > Dashboards ->Create Chart
Figure 1.
The empty chart would look as seen in figure 2.
Figure 2.
Insert or add a new source query as shown in figure 3. Metrics will be displayed as soon as valid query is set
Figure 3.
2. Intelligent alerting and monitoring
Customized alerts can be created and users notified via e-mail, slack, Pagerduty etc. Alerts can be further categorized into severity (info, smoke, warning, severe)
For more advanced options, explore the Wavefront documentaion on alerts.
In the example below, an alert is created for 100 % CPU usage for a container and an email target is set.
Navigate to alerts: Wavefront UI -> Alerts
Click on create alert as shown in figure 4.
Figure 4.
Write a time series query to generate 100% CPU usage by container, select type of warning and email target if any, as shown in figures 5 and 6 below.
Figure 5
Figure 6.
3. Sharing dashboards via URL with teams or colleagues (users who have Wavefront instance access) through slack, Pagerduty, Hipchat etc…
PKS dashboards can be shared among team members via generating a URL shown in figure 7. By accessing this URL, users can collaborate to view and discuss the intended metrics in detail.
It is secure, as the dashboards and chart contents are all transmitted via TLS.
Figure 7.
4. Granular and vibrant display of metrics
PKS metrics in Wavefront are more granular and immense data such as data received, data transferred, for namespaces, nodes and pods in a single graph with vibrant colors. All data can be displayed by hovering over the graph as depicted in figure 8
Figure 8.
Summary:
1.Wavefront provides monitoring of scaled container environments by enabling addition of new metrics without worrying about incremental costs. Customized dashboards can be created, edited and cloned with ease.
2. Apart from intelligent alerting mechanism in Wavefront, alerts can be tagged. Tagging can help to correlate container performance with host performance, saving time in problem isolation; using tags, several types of metrics can be aggregated and logical groupings can be created.
3. Query language can be customized and answers for respective parameters can be derived quickly.
4. By integrating Wavefront with DevOps toolsets such as PagerDuty, Hipchat and Slack, collaboration across several teams for dashboards and metrics can be enabled.