real-time cloud monitoring - signalfx · 2019-12-12 · real-time cloud monitoring for...

12
REAL-TIME CLOUD MONITORING for Infrastructure, Microservices, and Applications

Upload: others

Post on 25-May-2020

5 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

REAL-TIME CLOUD MONITORING for Infrastructure, Microservices, and Applications

Page 2: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 2

Operating in real-time is the imperative driving nearly every business. Customers want things now, not later, and they want their experiences to be error free. In our digital world, real-time, personalized engagement with mobile and web applications has become the foundation for customer experience, even in the most traditional businesses. For example, leading apparel companies use mobile apps to sell directly to consumers, often pushing flash sales of new items. Hundreds of thousands of users log in to these apps for the latest shoes or t-shirts, which can sell out in minutes. In this context, poorly performing applications caused by faulty code or inadequate capacity can result in massive losses in revenue and brand reputation. The ability to recognize potential issues and auto-remediate errors before customers are affected is a critical capability for every business.

This new world requires companies to rethink how they monitor their infrastructure, applications, and the performance of their business. More to the point, legacy monitoring solutions cannot provide the scalability, analytics, and automation capabilities needed to deal with the massive data volumes of cloud-native applications in real-time. That’s why companies are increasingly turning to SignalFx’s cloud monitoring solution with streaming analytics.

The Cloud-Native Journey

“Any loss of control by moving to the cloud has been easily offset by the convenience of using cloud infrastructure. ‘We get to focus on the things where we add value,’ — and that’s the goal of every company.” —Jason Fennell, SVP of engineering at Yelp

Every company follows its own unique cloud journey, but as those journeys progress, we see several distinct trends that are driving a new criteria for real-time cloud monitoring.

Page 3: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 3

INFRASTRUCTURE IS IN THE CLOUD AND CONTAINERIZED

The elasticity of the cloud and containers allows teams to rapidly provision and deprovision computing resources in order to meet demand or create cost savings. The ability to standardize application runtime environments with a single cloud provider or through container images also increases development velocity. Development organizations no longer need to make significant upfront investments in physical infrastructure.

OPEN-SOURCE SOFTWARE IS REPLACING PROPRIETARY APPLICATIONS

Though companies once relied heavily on proprietary solutions, free and open-source software has grown in popularity as more companies turn to developing their own applications. Choosing open-source components allows development teams a greater degree of flexibility and interoperability, while reducing upfront costs and the possibility of vendor lock-in.

APPLICATIONS ARE INCREASINGLY SPLIT INTO MICROSERVICES AND FUNCTIONS

Traditional applications combined functionality for multiple business processes in a single code base. Releasing a new version of an application could take years because of the need to fully account for complex internal dependencies. Microservices architectures break the application into a collection of independent services that can be individually developed and managed. This allows organizations to distribute the ownership, development, and operation of individual features to smaller, more agile teams that can create value more quickly.

DEVOPS PRACTICES LEVERAGING AUTOMATION AND CI/CD ARE PERVASIVE

Companies traditionally assigned responsibility for building and maintaining applications to separate development and operations teams - creating siloed organizations that struggled to coordinate efforts when troubleshooting issues. Today’s DevOps teams are pushing for continuous integration and software delivery, embracing a culture of collaboration to quickly release product innovations.

Page 4: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 4

The New Criteria for Real-Time Cloud Monitoring

native technologies and how well you monitor your entire stack. A new set of criteria come into play for real-time cloud monitoring.

FLEXIBLE, OPEN INSTRUMENTATION

In the cloud-native world, data collection is commoditized. It comes out-of-the-box using technologies such as service mesh, and is standardized by open, vendor-agnostic

An open standards approach to data collection provides transparency, better collaboration, and accurately captures developers’ instrumentation intent. In the modern, cloud-native world, SREs and development teams collaborate for observability.

REAL-TIME PROBLEM DETECTION

In today’s cloud-native environments, enterprises run microservices on hundreds and sometimes thousands of short-lived components, such as containers and serverless functions. Because every millisecond counts when it comes to delivering a flawless end-user experience, the ability to instantly discover and alert on meaningful problems in seconds is imperative.

DIRECTED TROUBLESHOOTING

Beyond catching problems as they occur with real-time detection, companies must also be

elastic and interconnected nature of architectures today create mountains of noise you have

direct the operator to the root cause of issues without alert storms.

BUILT FOR DEVOPS VELOCITY

To achieve DevOps agility and centralized control, companies need one tool that can enable the entire organization and individual developer teams with the relevant dashboards, collaboration tools, and self-service capabilities. These are critical capabilities for organizations embracing DevOps methodologies.

Page 5: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5

Traditional Tools Are Not Built for The New RulesTraditional monitoring tools were designed to address the needs of a small number of monolithic applications, which often were deployed in siloed environments and changed infrequently. Viewing different types of monitoring data (such as traces, metrics and logs) in isolation was not a problem. It was easy to identify and resolve issues because system behavior was largely predictable. These legacy monitoring tools were more concerned with data collection than aggregation and analytics.

The scale, elasticity, and churn of cloud-native environments greatly multiply the number of components that require monitoring. The adoption of microservices has shifted complexity away from individual lines of code to a collection of distributed systems.

With data collection becoming commoditized and microservices proliferating, the value of monitoring has shifted. Monitoring today is about aggregating data from unpredictable systems, then extracting the most meaningful signals from the vast array of noise. Legacy vendors are not equipped to meet the new criteria.

Accelerate Your Journey to Cloud-Native with SignalFxThe challenge of monitoring in a rapidly changing environment was first experienced by our founding team at Facebook in the mid 2000s. The development practices Facebook and other software companies embraced at the time have since become mainstream and are now more commonly referred to as DevOps. The original technical challenges our founders experienced at Facebook are now being faced by all companies digitally transforming their businesses.

SignalFx helps companies at any stage of the cloud-native journey separate meaningful signals from mountains of noise, using automation to address issues before they affect customers.

“With SignalFx we now have all the traces and metrics we need to immediately get to the root cause of any problem. That means more successful product launches— and happier customers.” — Adam Nutt, Engineering Manager at Nike

“For the vast amount of metrics we produced, SignalFx was the only solution we tried that didn’t fall over in the first week.” — Sam Eaton Director of Operations at Yelp

Page 6: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 6

monitoring for infrastructure, microservices, and applications. The platform discovers and collects metrics across every component in your cloud environment, and captures every transaction across

visualization, proactive alerting, and collaborative triage capabilities for today’s key monitoring use cases: Infrastructure Monitoring, Microservices Application Monitoring, and Digital Business Monitoring.

customer’s journey to cloud-native.

Page 7: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 7

architectures (e.g., VMs, OSS, containers, serverless functions, etc.) The patented SignalFlowTM real-timepredictive streaming analytics engine can process massive amounts of data in real-time, applying advanced

performance in aggregate - so you can know if a change is normal, an anomaly, part of a trend, or a threat

can detect and alert on meaningful conditions within seconds — and automatically remediate before

“We are now able to send and analyze metrics with four times more granularity than we had before, and predict outages on our customers’ instances of AWS before Amazon does.” — Aaron Pacheco, Product Manager for Acquia’s product delivery infrastructure

and Application Monitoring

a directed approach to troubleshooting. Unlike traditional APM solutions, µAPM analyzes end-to-end traces across service boundaries with open instrumentation options to provide flexibility for the end-

trace, and pinpoint trace anomalies in real-time. With deeper analysis than any other solution, the

performance of their services and being able to capture 100% of traces is very

‘hope we can find the problem’ because now we ‘know we will find the problem’.” — Trevor Rundell, Director of Engineering,

TM

TM

TM

Page 8: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 8

What a user wants to measure can be dependent on their role in an organization, and even what industry the company is in. For example, if you have a set of APIs that you expose to users, and have to maintain SLAs to specific customers for the uptime and performance of those APIs, you can monitor them with custom metrics. One leading retail footwear and apparel company measures the amount of shoes they sell per second during flash sales that only last minutes on their mobile application. Thanks

track high-cardinality custom metrics, because ultimately, it’s about monitoring the KPIs that matter most to their business.

customizable depending on what matters most.

“Analytics is really important for us because it lets us make sense of all the various

challenge, to understand how our customers deal with very large models and interact with upstream and downstream providers around the world.”

— John Rousseau, Director of Technical Operations at Onshape.

with Streaming Analytics and NoSample Tail-Based Distributed Tracing

distributed traces in one unified platform. That means you get open instrumentation, real-time problem detection, and directed troubleshooting packaged up as a service for today’s DevOps methodologies.

COLLECT DATA FROM ANY SYSTEM WITH FLEXIBLE AND OPEN INSTRUMENTATION

metrics for over 150 cloud-native technologies – as well as out-of-the-box trace collection from dozens of the most popular open-source frameworks and libraries.

without vendor lock-in. Lightweight and easy to set up, the Smart Agent provides automatic discovery, configuration, and instrumentation of metrics and traces for your most critical environments,

the normal behaviour of a system and identifying the anomalies in real-time.

Page 9: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 9

REAL-TIME PROBLEM DETECTION BASED ON STREAMING ANALYTICS

SignalFx is able to alert on any condition from infrastructure, application, and microservices in seconds with the SignalFlow Streaming Analytics Engine. The architecture that powers SignalFlow is different from other vendors in the monitoring space because it’s based on stream processing as opposed to batch querying of metrics. Instead of than having to query the database with the arrival of every new datapoint, SignalFx Metadata Router and SignalFx Time Series Router both have a pub/sub bus that provides an event-driven approach. By analyzing the data and matching it with any analytics jobs currently running, SignalFx is able to analyze and fire alerts within seconds versus the minutes it takes for batch querying systems.

SignalFlow also allows powerful analytics to be packaged in an easy-to-use interface. With statistical functions and algorithms embedded into charts and alert templates, users can alert on dynamic thresholds, detect outliers, measure sudden changes, and track historical anomalies.

If you want to unlock even more power from SignalFlow, you can also access the SignalFlow API to customize your data streams. SignalFx provides a Python-like query language that is powerful enough for developers to programmatically create monitoring content and define analytics, but still easy to learn.

Page 10: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 10

DIRECTED TROUBLESHOOTING WITH NOSAMPLE TAIL-BASED DISTRIBUTED TRACING

Beyond the ability to gain full coverage for metrics across your infrastructure, application and business,

NoSample ™ tail-based Distributed Tracing architecture that doesn’t force you to naively sample data. By

pattern recognition, you’re able to quickly sort through the noise and pinpoint the root cause of the

and outlier traces.

Page 11: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 11

MONITORING AND ANALYTICS AS A SERVICE BUILT FOR DEVOPS VELOCITY

Bureau provides this customization with centralized control over usage, access, and permissions. With

longer have to worry about surprise fees and overages. These are critical capabilities for organizations embracing DevOps methodologies.

Conclusion

changed almost everything about how businesses compete today. It has created a need for a new generation of cloud monitoring solutions that can easily ingest data from distributed, ephemeral environments, while providing powerful real-time analytics to identify, troubleshoot, and remediate

and expect.

applications. The platform discovers and collects metrics across every component in your cloud environment, replacing traditional point tools and providing real-time predictive analytics. The

applications, It provides powerful visualization, proactive alerting, and collaborative triage capabilities

consumer products, retail, communications, media, entertainment, and web-scale players like Yelp,

Ventures, and General Catalyst.

Page 12: REAL-TIME CLOUD MONITORING - SignalFx · 2019-12-12 · Real-Time Cloud Monitoring for Infrastructure, Microservices, and Applications 5 Traditional Tools Are Not Built for The New

SignalFx is the only real-time cloud monitoring platform for infrastructure, microservices, and appli-cations. The platform discovers and collects metrics across every component in your cloud environ-ment, replacing traditional point tools and providing real-time predictive analytics. The massively scalable SignalFx architecture is optimized for cloud infrastructure and microservices-based applica-tions and provides powerful visualization, proactive alerting, and collaborative triage capabilities for organizations of all sizes. SignalFx is used by leading enterprises across high tech, financial services, consumer products, retail, communications, media, entertainment and web-scale players like Yelp, HubSpot, Acquia, and Kayak. SignalFx is venture-funded by Andreessen Horowitz, Charles River Ventures, and General Catalyst.

©2018 SignalFx, Inc. All rights reserved.