This presentation was given at TopConf Tallinn in May 2015.
Title: Driving the DevOps feedback loop in the Enterprise
Description (brief overview for marketing purposes, max. length 400 characters-about 65 words):
DevOps requires feedback loops backed by consistent data, within enterprises many layers of tools make visibility a challenge. This session outlines modernizing monitoring and providing the right data collection capabilities to support agile application operations. This session will outline the use of open source and commercial software solutions which fit startups to the enterprise.
Abstract (Longer, more detailed description (3-6 paragraphs, bullet points welcome) of your presentation to help the program committee understand what you will cover. Please keep in mind that if your proposal is chosen, this abstract will also appear on the website to help conference attendees decide if it's right for them.):
Enterprises do not have the benefit of starting from scratch, they must implement DevOps in an existing environment, often managing heritage investments along with fast moving projects. The monitoring tools unfortunately fall into the same category. These complex, dated, and costly monitoring tools don’t meet today’s needs in providing the visibility required for agile development that leverage continuous delivery and DevOps. Not only are the tools an issue, but the visibility and approaches are also problematic.
The use of open source tools has been the de facto approach for DevOps, but this approach comes with it’s own set of challenges in terms of managing a complex stack of tools with varying quality of support/community. The move to a metrics-driven feedback loop enables teams to act fast at any stage of the product lifecycle. We will outline common technology stacks and approaches using graphite, statsd, and collectd along with nagios, munin, and other system monitoring tools. We will also outline commercial solutions, and how these tools tend to fall into silos as well.
Monitoring must be reformed with new platforms and technologies, fitting specific design patterns which meet the needs of DevOps teams, these will be outlined and explained from a requirements perspective, providing a roadmap for how to apply these learnings inside your company. Join this session to take back an evolved approach to unified monitoring and how you can get your company to be metrics driven
Attendees of this session will walk away with a clear understanding of:
Understanding today’s tooling limitations
Open source tool sprawl issues
Fundamental patterns of monitoring
How these will be solved in the future