appsphere 15 - achieving enterprise agility, superior scalability and deployment at speed with...
TRANSCRIPT
Achieving Enterprise Agility, Superior Scalability and Deployment at Speed with APMaaS Oscar García – NTT Europe COO / Cloud & Managed Services Business Unit
Agenda
The Presenter Pain Points – Customer Needs Our Approach
– Deployment at Speed – Smart Autonomics – Deep APMaaS
Putting it all together – Some real-world examples
A glimpse of what I will be covering
The Presenter
Oscar García • I run operations for Cloud & Managed Services, part of
NTT Communications • Largest Telco globally = big customers = slow???
• I sold them my company (second!) • I am all for agility, speed and operational excellence on a budget
• I love AppDynamics. Since day 1.
PAIN POINTS – CUSTOMER NEEDS What we are trying to fix
Copyright © 2015 AppDynamics. All rights reserved.
4
Pain Points
• Lack visibility of the application state
• Complex, slow and inconsistent deployment of tooling
• 24x7 NOC availability • Use of internal ops teams • RCA analysis and expertise
Visibilty
Deployment at scale
24x7 Alert management (with Context)
Operations agility
Access to SME's
DEPLOYMENT AT SPEED Agile and consistent delivery
Copyright © 2015 AppDynamics. All rights reserved.
6
Classic Monitoring challenges
• Nodes need to be registered at the poller • Packaging is complex with dependencies • Inconsistent and partial coverage risk • Inbound network connectivity needs • Not designed with the cloud in mind
Our Deployment approach
• Let’s leverage AppDynamics features fully • Engage with our Professional Services to:
Copyright © 2015 AppDynamics. All rights reserved. 8
Consult and discover
• Workshop with the customer on the solution
• Understand needs and constraints
• Capture requirements
Design the architecture
• Controller connectivity
• Network architecture
• Application and layer grouping
• Potential operations automations
• Optimisations
Build the solution
• Build the central repository
• Prepare network connectivity
• Prepare packages
Deploy the solution
• Push the packages • See how the agents begin to automagically appear...
Continuously support the solution
• Develop and implement consistency checks
• Audit for full coverage
• Refine the process / optimise
Benefits
• Our deployments are: – Scalable and Secure – Ensure 100% coverage – As we are Application savvy, we know how to integrate
– No relevant part of the solution is left un-monitored… ever – Our operations teams are fully up to speed with the solution
Copyright © 2015 AppDynamics. All rights reserved. 9
Automation in Operations (classic!)
Focus on incident resolution…
Something fails
Automation Engine
captures the event
Automation Engine
fixes (99% = restart)
Service is recovered
Creates more problems than it solves – No Root Cause Analysis so it will repeat, again and again, again… – The platform becomes dependent on the automation engine
Operations (thoughtful!)
Focus on RCA, and analysis
Copyright © 2015 AppDynamics. All rights reserved. 12
Something fails
Monitoring Engine
captures the event
Automation Engine
escalates to the on-call
team
Team investigates/
escalates further
Somebody fixes (99% =
restart) Issue is
fixed
Creates more problems than it solves – Affects service as it does not provide a quick fix – It is expensive, as SME’s need to be on-call
Smart Operations Autonomics (Our way)
Copyright © 2015 AppDynamics. All rights reserved. 13
Something fails
Automation Engine captures the event
Automation Engine fixes (99% = restart) Service is restored
AE triggers a Snapshot in AppD
Information is sent to SME's in BH
Network SME analyses
Cloud SME Analyses
DB SME analyses
App SME analyses
Benefits
• Remediate quickly – Service is not affected – Business is happy
• Provide context – Capture all background info
• Look for RCA – In BH = with all the bright minds
• Fix permanently – Avoid restarts and dependency on automated remediation – Sustainable and healthy application
Copyright © 2015 AppDynamics. All rights reserved. 14
DEEP APPAAS Taking advantage of SME’s as a Service
Copyright © 2015 AppDynamics. All rights reserved.
15
Deep APMaaS – where are we
APM is disruptive • Full end-to-end visibility • Deep visibility • Quick Root Cause Analysis and APMaaS even more • Fast and agile deployment • Easy alignment of cost and revenue • Immediate return on investment
Copyright © 2015 AppDynamics. All rights reserved. 16
Deep APMaaS – why?
With all the goodness and insight of APMaaS, we need to: • Triage and use that information • Make sense and analyse all the outputs
Copyright © 2015 AppDynamics. All rights reserved. 17
Our approach for Deep APMaaS
Copyright © 2015 AppDynamics. All rights reserved. 18
FrontEnd
• Provided by us
• Smart Engineers
• AppD savvy • Always there
Core Operation
• Customer Owned
• Embedded in the dev process
SME's
• Deep AppD knowledge
• Deep Tech expects
• Asynchronous
Deep APMaaS - Front-end
• Redundant 24x7x365 Operation Solution – Fully secure and ITIL compliant processes (ISO27001 / ISO20000)
• AppDynamics Savvy – Trained in AppDynamics and with at least 1 year of experience – Not only forwarding alerts, able to retest and add brains!
• Supported by Technical Service Managers – Own the Problem management process – Own the Continuous Service Improvement process
• No access to the managed solution
Copyright © 2015 AppDynamics. All rights reserved. 19
Deep APMaaS – Core Ops
• Owned by the customer (optionally) – Development cycles (specially DevOps-ey) demand agility – Embedding a 3rd party in the process is tough
• The Front-end feeds Core Ops – Curated alerts, converted in to incidents
• We integrate with pretty much anything – No hassle, just us embedded in the customer OSS
• Do you want us to take-on the core ops? No problem!
Copyright © 2015 AppDynamics. All rights reserved. 20
Deep APMaaS – SME
• Deep technical expertise to support core operations • Escalation point for the Front-End or Core-Ops for tackling
complex Applications issues – Root Cause Analysis – Issue Resolution – Automation implementation – Solution Optimisation
Copyright © 2015 AppDynamics. All rights reserved. 21
Deep APMaaS – SME Coverage and Benefits
• Core technologies – AppDynamics – Cloud Technologies – Networking – Storage – Compute/OS – Databases – Application Servers – Web Servers
• Benefits – Access to experts for RCA – Asynchronous operation – Vested interest in fixing
issues – Peace of mind!
Copyright © 2015 AppDynamics. All rights reserved. 22
Standard use case
• Running on Hyperscale cloud
• Using AppDynamics SaaS controllers
• Mix of technologies
• Customer provided core operation
Deployment at speed
• Consulting • Designing • Deploying
Smart Operations Autonomics
• 24x7 NOC support
• Auto remediation • Problem Management
Core Operation
• Customer owned • Close to the cycle
Deep APMaaS
• SME involvement
• Complex RCA • Complex Fixes
Example – Online Hotel Booking System
• Challenges – Changing environment – Very seasonal and peaky
traffic – Response time = revenue
• Environment – More than 20 AppServers, 10
DB servers, Cloud – Auto scaling for AppServers
• Solution – Full APMaaS service
• Benefits – End to end deployment time
reduced from 4h to 5m – FCIR increased from 50% to
90% – Full coverage = no lost
revenue – Access to Oracle, Java SME’s
Copyright © 2015 AppDynamics. All rights reserved. 25
Example – Massive Global Retailer
• Challenges – Lack of consistency with
departmental IT state – Stringent security needs – Demanding availability reqs
• Environment – Mix of Windows/SQL and Linux/
MySQL environments – On Hyperscale and physical
• Solution – Full APMaaS service – Automation & Bootstrapping – Integration with CloudFormation
• Benefits – Deployment time reduced from 5h
to 5m – FCIR rate increased from 30% to
70% – InfoSec team happy / auditing – Consistent visibility
Copyright © 2015 AppDynamics. All rights reserved. 26
Example – Leading Manufacturer of Print / Cameras • Challenges
– Time to market for solutions – Extremely peaky traffic – Lacking Application quality – Response time = impact
• Environment – Templated environments – Auto scaling in some cases – Both Public and Private
• Solution – Full APMaaS service – Automation & Bootstrapping – Packaging of Apps for Auto
scaling
• Benefits – End to end deployment time
reduced from 4h to 5m – FCIR rate increased to 80% (from
40%) – IR time decreased from 60m to
10m
Copyright © 2015 AppDynamics. All rights reserved. 27