On MultiCloud Management. Q&A with Chris Psaltis
Q1. Mist.io is the creator of Mist. What is it?
Mist is an open source multicloud management platform. It supports all popular infrastructure technologies including public clouds, private clouds, hypervisors, containers and bare metal servers. It provides a unified interface for performing common management tasks like provisioning, orchestration, monitoring, automation and cost analysis. It comes with a RESTful API so you can easily integrate it into your existing workflows. Our customers reduce server bills and management overheads by 60%, while increasing their velocity by an order of magnitude.
Q2. What are the main challenges in managing a mix of public and private clouds?
In heterogeneous environments, visibility, uniform processes and centralized control are all hard problems to tackle. What resources do we have? How do we integrate all these APIs and tools? How can we provide self-service workflows? Can we optimize usage and cost? Unless you can confidently answer all these questions, you are probably moving slow and leaving money on the table.
Q3. How is Mist helping on this?
Mist integrates with dozens of point solutions and offers a single pane of glass to help you answer questions like the ones above. A good example is Mist’s tight integration with Telegraf and InfluxDB. You can begin collecting monitoring metrics right away and evaluate them in a backdrop of inventory and cost information. This way, you are not missing context and you can make better decisions faster.
Q4. What are the main advantages of the integration of Mist with Telegraf?
First of all, Telegraf is really easy to install. It’s written in Go and you get just one binary which you can run without other dependencies. This is especially important to us because we can automatically deploy Telegraf on any operating system our users have. Before Telegraf, we were using CollectD. Back then, we had to juggle with links to system libraries and it was practically impossible to support more than a handful of distros and versions.
Telegraf being open source is also a big plus for us. We are deploying a vanilla upstream version of the agent, and we can be very transparent about what code is running on our users’ servers. Comparing Telegraf to CollectD, Telegraf is better maintained and has more frequent releases. If you check the respective GitHub repos you will see that Telegraf is better across all metrics, e.g. number of commits, contributors etc.
Finally, with Telegraf we can tap in a rich ecosystem of plugins and integrations. This helps us offer a lot of functionality without reinventing the wheel.
Q5. Why do you use InfluxDB? And for what?
We are using InfluxDB for storing monitoring metrics and metering data, e.g. how long has server X been running.
We love InfluxDB because of the flexibility it offers. In earlier versions of Mist, we were depending on Graphite and we had to preconfigure the granularity of ingested metrics AND preallocate the required disk space. With InfluxDB we don’t have to worry about any of that.
The licensing is also great. InfluxDB has a real open source license, as in MIT. In contrast to code available alternatives, the MIT license gives us the freedom and peace of mind we were looking for.
Finally, the popularity of InfluxDB is helping us avoid friction with our users. Chances are that they have experience with InfluxDB, and they might even be running it already. In these cases, we can just connect to their existing cluster, and we don’t deploy yet another DB instance.
Q6. What kind of monitoring metrics do you offer?
Right out-of-the-box you can monitor all usual system metrics like CPU usage, load average etc. Also, Telegraf gives us access to a rich library of application monitoring plugins to gain further insights into how they perform. Finally, for cases where you need to monitor something more custom e.g. ping times to an IP, a value from an internal DB etc., you can deploy custom metrics through the web using Telegraf’s Python plugin.
Q7. Who are your clients? What do they normally use Mist for?
Our customers include technology companies like SevOne, which develops network monitoring software, and Windstream Telecommunications which offers unified-communications-aaS. We also see traction in other tech-intensive verticals like financial services (National Bank of Greece), retail (Shoprite, the largest supermarket chain in Africa) and education (HPC group of the University of Boulder).
Mist helps them implement self-service workflows in a well-governed way. It also offers them visibility into how systems are utilized and what optimizations are needed. You can check out a real-world example from SevOne on our blog).
Q8. Since Mist is open source, how can developers contribute?
Mist is licensed under the Apache v2, and we love working with our community members. If you’d like to contribute to our project, head over to https://github.com/mistio/mist-ce and look for a “good first issue”. For more help, you can open a ticket or reach out to us directly at firstname.lastname@example.org.
Chris Psaltis, CEO & co-founder at Mist.io.
Chris Psaltis is the co-founder and CEO of Mist.io, an open source multicloud management platform that gives organizations an off-the-shelf way to make multicloud manageable. His background is in computer vision and machine learning. He loves open source, Python and solving problems. Before founding Mist.io, Chris worked as a researcher, software engineer and project manager.
Sponsored by InfluxData.