The Internet is a massive labyrinth of connections and intersections. Without a map or a guide, it would be almost impossible to find anything inside of it. That's why we need a translator to help us understand DNS, the "Internet's phone book." For many organizations, websites, and companies, that translator is Dyn, the market's leading Internet Performance Management (IPM) solution.
Dyn collects over 200 billion data points everyday, in order to analyze online traffic, spot bottlenecks, and make real-time routing and infrastructure decisions. Trusted brands like Zappos use Dyn’s IPM solution, which offers “dynamic steering” of their Internet traffic, maintaining the high availability and performance required to meet expectations for outstanding customer experience.
Given Dyn’s customers' demanding expectations and Dyn's data-intensive solution, Dyn's systems need to run at peak efficiency. Staying true to their roots in high performance analytics, they've implemented performance monitoring tools for the apps and databases that support their solutions, so their team can quickly spot and resolve performance problems. This is no small task given their data-intensive business.
Dealing with a Complex Environment
Dyn improves Internet performance by enhancing customers’ understanding of the Internet's ever-changing environment, and keeping them aware of the various options for reacting to changes in web traffic.
Additionally, Dyn and its customers have to meet the challenge of making changes in real time, in active production systems. As in dynamic fault management—where systems can’t be taken down for repair, but must fixed on the fly—every adjustment must be accompanied by live feedback, to confirm that the problem is resolved.
Dyn accomplishes this by steering traffic and monitoring results—it's a constant search for the most efficient path. The objective is not to shoot for a perfectly tuned system—which is pretty much impossible—but to respond as quickly and effectively as possible, while still operating in a highly complex environment.
Improving Time to Resolution
Cutting time to resolution a key goal for any engineering team. So how does Dyn do it? VividCortex has played a role. Dyn's engineering team was nice enough to share with us how database monitoring has helped them improve their time to resolution for key issues. The following were some of the key points from our discussion:
- Monitoring highly distributed services. Running highly distributed services with hundreds of database instances presents a major management challenge. Before VividCortex, Dyn needed a tool that could monitor every one of their database instances; aggregate performance data to quickly isolate issues; and create custom alerts. The ability of a monitoring solution to go broad and then deep was critical to Dyn for improving resolution time.
- Analyzing using side by side comparisons. Without the proper analytic tools, isolating the cause of slow performance can be tiring and stressful. In instances where Dyn noticed high wait states, with processes stacking up and overwhelming connections, they were able to open their database metrics and compare graphs side by side. By visualizing large data sets, it became much easier to determine which issues were related or causal.
- Communicating and collaborating. The Dyn team has a perpetual “need for speed.” They regularly communicate in real time, using social tools like Slack to keep the right team members informed of key developments. Sometimes, however, email or a chat service isn't enough. If the database presents an unusual development, Dyn's data engineers need a way to share full views of their exact dashboards. They use VividCortex's deep-linking capability to send graphs and charts with colleagues, simply copying the links directly into their chats and emails.
- Learning and improving. With VividCortex, Dyn's team has the option to conduct post-mortems with precise database performance data. This not only helps build the knowledge base of the team with respect to potential issues and their causes, but it helps them become more adept in the use of the analytics tools and interpretation of what they uncover. The whole team's understanding of their systems continues to grow, as they can all review what their databases were up to, after the fact.
Happier Customers... and More Sleep
Dyn meets the high standards of their customers thanks to their deep understanding of "high-performance." With VividCortex, they can apply that understanding not only to improving the IPM space for their clients, but also to their own database systems.
How does the Dyn feel about their database analytics tools? A member of Dyn's team asked us,“If I get paged at 2 am, do I want to be up for a half hour or three hours?” Without a database monitoring solution, a long night is ahead. With VividCortex, Dyn's engineering team can quickly spot and resolve issues. And get more sleep.