Grafana / GrafanaCon EU 2018

Significant amounts of time have been invested in performance testing and analysis of OpenStack.

In this talk I share how we use Grafana and many other Open Source tools to monitor the performance of an OpenStack Cloud while benchmarking.

We'll show the components of the monitoring stack and benchmarking tools, share the dashboards which help speed up OpenStack Performance Analysis, and run through scenarios showcasing how the dashboards have exposed issues and helped improve OpenStack performance.

Jsonnet is a tool that allows you to write Json files easily, handle repetitions, data, etc.

Terraform is a tool that works with may API's to create/update/delete objects from a codebase. It is mainly known in the cloud area but has support for grafana too.

Combining those two solutions allows you to write many dashboards in no time - in a reproducible way. And even to create basebricks for multiple dashboards and reuse them at will.

Since 2013 New City Energy has been helping the City of D.C. measure and improve the efficiency and performance of its building portfolio through a number of policy and technology innovations. In 2016, NCE began building out a cloud based platform to allow deep insight to the portfolio at a room-by-room level.

Learn how we've structured the architecture to enable this platform, the context of how Grafana interacts with the other components in the system, and how the entire system delivers value to the City.

This talk outlines the components of our monitoring stack (what we're allowed to share) and what drove us to seek out Grafana in the first place. We discuss our starting point, the challenges we face on a daily basis, and what we plan for the future as we continue to scale.

Andrejs gives insights of how visualization of data can improve testing process by Grafana usage in performance tests, monitoring of test environments and other QA activities and which time series data base fits best for such tasks.

Grafana has a great UI for creating dashboards, but what happens when you want to start implementing common patterns, or change your dashboard at the same time that you change your monitoring?

Why not treat your dashboards like code? We do that with grafanalib, which is a Python EDSL for building Grafana dashboards.

This talk will show how you can use grafanalib to manage your dashboards, some cool things you can do with it, and talk a little bit about its design & internals.

This is a story about dealing with metrics at scale. A lot of metrics.



This is our story of the challenges we’ve faced at Booking.com and how we made our Graphite system handle millions of metrics per second.

You should expect to learn about one of the most high load Graphite stacks, the problems it poses and the challenges in maintaining it and scaling it further, pushing Graphite to its limits and beyond.

Dave just started his new job as a Systems Engineer in this strange IT company. His ambition is to become the most aspiring and omniscient operator the company has seen.

Through his journey of debugging issues and hunting users he learns many things helping him to troubleshoot and improve running systems. Unfortunately, at some point in his career, Dave reaches a moment where none of the learned techniques can't help him

. What he needs is something that that can display all the data he collects with different tools in one single place.

This story is about how the art of data visualisation can change your whole perspective and help you find the right answers

Amgen has been on a journey to confront the multi-million dollar elephant in the room. Technical debt that had been passed down for 10 years in their monitoring portfolio.

But how to overcome inertia and a failure to understand how to maximize internal value for data that belongs to organization to begin with?

I'll discuss lessons learned, where it was, where it is now, and what steps we took to achieve technical value given back to the business through open source, modular layers and with a design for programmatic monitoring at global enterprise scale.

GrafanaCloud makes it easy to set up, operate, and scale a complete Grafana stack in the cloud.



We'll discuss the components of GrafanaCloud, how it scales, and how to achieve true interoperability leveraging open source software and avoid SaaS lock-in.

Fred Moyer talks about IronDB, and the new Grafana data source they developed.

AI platform empowers eBay’s data science and engineering teams to experiment, productize and enable machine learning at scale. Organizations within eBay are allocated fixed capacity and knowing resource utilization over time enables efficient capacity allocation. In addition, user's would like to know in real time resource usage of their jobs.

Grafana dashboards provide an organization view and user view of resource utilization. In addition, it allows user's to search for specific job identified by unique ID displaying resource utilization of each task within a job.

After insourcing the majority of subsurface technical storage at Shell uptream, we were limited by the commercially available monitoring solution. A Proof of Concept was held using a data collector, InfluxDB and Grafana. This enabled us to monitor the full globally installed base on a single dashboard as well as drilling deep into specific storage clusters.

We'll show the interesting and completely unexpected advantages of implementing Grafana and the insights we've gained.

Carl Levine discusses the NS1 platform and why gaining visinility into your DNS is critical.

Natel Energy is building a low cost, distributed, grid-scale hydropower turbine. We began using traditional industrial control packages to manage and develop our projects. These products tend to be expensive, fragile, hard to scale and ugly. With Grafana, we found a great platform to build a good looking, secure and scalable control system. Rather than build a new application we are now building all our operational interfaces directly in Grafana plugins.

Representatives of 4 TSDBs sit down with Grafana Labs co-founder and CEO Raj Dutt to discuss different points of view about time series data, query language and more.

Building a company around an open source project is not for the faint of heart. Find out what we've been up to since our last GrafanaCon, How the team has grown, our view on the open source software landscape, and what's to come for Grafana Labs.

The new InfluxDB query language is functional in nature. Conceptually it looks like D3 or jQuery. Functions are chained off each other. This gives greater flexibility for what the language can do and opens up new possibilities for UI builders. This talk will cover the details of the language and user interface ideas that can pair up nicely with it. We'll look at query examples that perform math on metrics, interpolation, and similarity measures.

In this talk, Matt Toback discusses the posters that Grafana Labs has been designing for events over the past few years and unveils the latest in the series at GrafanaCon EU.

The launch of Grafana v5 is probably the single biggest enhancement of the project's history. New features include a totally revamped layout engine, dashboard folders, and some foundational UX/UI improvements.

We'll dive into the new features as well as what's on the horizon for Grafana

The RED Method defines three key metrics you should measure for every microservice in your architecture. It gives developers a template for instrumenting their services and building dashboards in a consistent, repeatable fashion. We will discuss patterns of application instrumentation, where and when they are applicable, and how they can be implemented with Prometheus. We’ll cover Google’s Four Golden Signals, the RED Method, the USE Method, and Dye Testing. We’ll also discuss why consistency is an important approach for reducing cognitive load, and discuss the limitations of these approaches and what can be done to overcome them.

From zero to Kubernetes: Which challenges do I face while migrating a cloud-unready Java application such as HiveMQ from a classical Linux IT- environment to Docker-Orchestrator Kubernetes?

In this talk we want to show you the technical problems we had to solve and how monitoring setup helped us to handle critical incidents in production.

In this talk, we take a somewhat heretical stance in the monitoring world, and describe why and how we chose to use PostgreSQL as a Prometheus backend to support complex queries (and get a full SQL interface).

Until recently, the databases supported by Grafana have been NoSQL systems, offering SQL-like or custom query languages that were limited in scope, and designed for specific data model and architectures in mind.

In this lightning talk Erwin explains how he gained insights in his power usage using a raspberry pi zero with Prometheus and Grafana.

How Grafana can be used to visualize future data, although default usage is focussed on historical data.

In addition to this, they'd like to give a short introduction on the increasing weather impact on financial markets like the energy trading business and use Grafana to demonstrate how weather forecasts, resulting renewable energy production and finally price developments are connected.

Finally, a discussion, on how Grafana could be improved for the usage in non-monitoring businesses.

Monitoring Kubernetes is vital to understanding the health and performance of a cluster, but which metrics are most important to add to your dashboards and alert on?

Jacob will discuss how to most effectively monitor and visualize your Kubernetes cluster using the Grafana Kubernetes plugin and PromQL.

Percona has integrated Grafana + Prometheus to use as foundation for their Open Source Database Monitoring Solution - Percona Monitoring and Management, which now has thousands of active deployments.

In this talk Dimitri will talk about their experience integrating Grafana with Prometheus and deploying it at scales - the obstacles they had to overcome and exciting possibilities Grafana has offered Percona.

Dan Cech discusses some of the new features in Graphite.

Tom Wilkie covers the Prometheus project and what's new.

Erik Nordstrom provides an update of TimescaleDB.

When performance metrics are combined with active user experience and machine learning, we can gain new insights about why a business critical application is slow, what components are involved, and how this impacts users and hardware.

By interlinking several levels of dashboards, users from different departments can get a more transparent view of the issue and exchange problem related information as easy as sending a hyperlink to a specific dashboard to their colleagues. Thus troubleshooting and root case detection become much easier.

But collecting up to 500 million data points per day per customer is not trivial.

Torkel Odegaard kicks of GrafanaCon EU in Amsterdam, March 1, 2018.