About
Klusterbud
KlusterBud makes PromQL queries to the user's running instance of prometheus, collecting metrics from nodes and comparing them to other nodes within the same replica set.
When an anomaly is found, the identifying information of the node is served back to the user.
Users can choose from a dropdown list of replica sets in their cluster, and start watching for any anomalies.
How it
Works
Our priority when creating KlusterBud was its ease of use and pain-free implementation into any developer’s workflow. A React framework allows for the streamlined, easily navigable front end necessary for this vision without compromising on features or performance.
In the back end, KlusterBud sends PromQL requests to the user’s instance of Prometheus, gathering specific information about the performance of the nodes and their replica sets.
Technologies
Here are some of the technologies we have implemented in Klusterbud, feel free to click on the logo, we have linked the official docs to have a better understanding on them.
Features
KlusterBud offers a unique, straightforward interface for finding anomalies in nodes’ performances.
First, the user selects the replica set that they wish to examine. Then, they click ‘start watching.’
- Click the logos to check official docs
Kubernetes
Docker
Prometheus
React
Node & Express
That’s it. If an outlier in cpu usage emerges for one of the nodes within the replica set, key metrics and identifying information of that node will be served to the user.
KlusterBud’s
Future
There is still so much we will love to add, but here are what we are planning to work on next:.
Containerization
A containerized version of the application is the next step to reach a massive amount of new features, improvements, and optimizations.
These include:
- Laying a foundation that allows all additional features to not compromise the streamlined setup and use
- Removing the need for the user to have an existing and exposed instance of prometheus running on their cluster, greatly improving ease of setup
- Closer monitoring as a result of deploying KlusterBud directly into the cluster
- With KlusterBud running as a fully-integrated part of a developer’s cluster, we could create real-time constant monitoring features for a production workflow. This could include a variety of metrics more useful for a production workflow, as well as alerts for node anomalies and failures.
Meet the team
Michael Wang
Fredy Rubio
Christopher Lu
Asher Mitchell-Smith