Originally published on September 9, 2019, at ObjectRocket.com/blog
Our mission at Rackspace ObjectRocket is to put our customers’ databases on autopilot, so you don’t have to worry about your data and focus on building great apps. However, even if your database is on autopilot, maintaining system observability is still an important goal.
Maintaining an accurate picture of what’s going on with your data store and what impact your latest application changes are having on the performance of your database is incredibly important.
That’s why we launched our brand new metrics dashboards on our CockroachDB®, PostgreSQL®, and Elasticsearch® products on our new platform!
What you get
The first thing you’ll notice when you log into ObjectRocket’s Mission Control is that we have a Grafana® Metrics option in the main menu.
Click the Grafana Metrics icon to open a new browser window and log into a hosted Grafana server by using single-sign-on, where you can see metrics dashboards for all of your instances on our new ObjectRocket platform.
With Grafana, as part of your Rackspace ObjectRocket subscription, you get:
- Single-sign-on between Mission Control and Grafana
- Dashboards with key metrics for all of your ObjectRocket instances
- All your metrics in one place
- Unlimited metrics retention
What better way to appreciate the value of these metrics than to see it for yourself! So, go sign up at ObjectRocket, create an instance, and check it out today! To learn a little more about how we did it and what we’re looking at next, read on.
How we did it
At ObjectRocket, we’ve always leaned on comprehensive metrics to help us manage and maintain our customers’ data stores. For those metrics, we’ve been heavy users of Prometheus. Every instance that we manage exports metrics to local Prometheus servers that our support teams use to monitor and diagnose datastore issues.
Though we have a wealth of data internally, our customers commonly request access to the same metrics we use. When we started building our new platform, we set out to deliver on that request. To get there, we solved a few problems that Prometheus doesn’t: retention, global queries, and multi-tenancy.
To solve retention and global queries, we turned to the open-source project, Thanos. Thanos enables you to query across our global fleet of Prometheus servers by using a single endpoint, compact our older data to manage storage, and provide unlimited retention. It was the ideal solution, and it’s worked perfectly.
Multi-tenancy posed a different challenge. Grafana, our preferred front-end for dashboards, supports multi-tenancy out of the box, but Prometheus doesn’t. We developed a solution to fill that gap. We call it prometheus-filter-proxy, and it allows us to filter only the data that belongs to the user performing the query.
Along with our platform, we have all the pieces we need to provide metrics to our customers. Though we just touched on the overall architecture described in this post, we’re planning follow-up content that goes into a little more depth on the architecture, challenges we faced, and decisions we made along the way. Stay tuned.
What’s next
This metrics feature addresses a huge request from our customers, but there’s still more on the way! Another feature is a subset of these metrics displayed in our new control panel, Mission Control. Though you’ll go to Grafana to do deeper metrics analysis, we’re putting the top metrics in Mission Control so you can get a quick overview of your instances. We’re also adding additional configurability and dashboard controls in Grafana. So, once again, sign up and check it out today!
Use the Feedback tab to make any comments or ask questions. You can also click Sales Chat to chat now and start the conversation.