SlideShare a Scribd company logo
Measured	
  Availability	
  and	
  
Performance	
  
	
  
	
  
Sanjay	
  Singh	
  
singh.sanjay@yahoo.com	
  
Agenda	
  
•  FCI	
  –	
  Failed	
  Customer	
  Interac@on	
  
•  TP	
  –	
  Top	
  Percen@le	
  
•  RUM	
  
•  RUM	
  with	
  SPA	
  
Availability	
  and	
  Down@me	
  
FCI	
  
•  FCI	
  is	
  Failed	
  Customer	
  Interac@on	
  which	
  is	
  
measured	
  for	
  a	
  par@cular	
  page/workflow.	
  
– A	
  customer	
  interac@on	
  can	
  be	
  page	
  view,	
  form	
  
submission,	
  etc.	
  In	
  HTTP	
  terms,	
  customer’s	
  
interac@ons	
  are	
  GET/POST	
  request.	
  
– If	
  the	
  page	
  response	
  is	
  not	
  as	
  per	
  the	
  expected	
  
behavior	
  of	
  the	
  applica@on,	
  then	
  it’s	
  a	
  failed	
  
customer	
  interac@on.	
  
How	
  is	
  FCI	
  calculated	
  ?	
  
•  FCI	
  is	
  calculated	
  on	
  per	
  page	
  basis.	
  If	
  page	
  
execu@on	
  is	
  not	
  as	
  per	
  the	
  expected	
  behavior,	
  
then	
  it	
  is	
  counted	
  as	
  failure.	
  
•  Iden@fy	
  all	
  failure	
  cases	
  and	
  log	
  it	
  
appropriately	
  on	
  the	
  server	
  with	
  user	
  session’s	
  
info.	
  
Performance	
  
•  As	
  per	
  Google	
  –	
  
•  Increasing	
  web	
  search	
  latency	
  100	
  to	
  400	
  ms	
  reduces	
  the	
  daily	
  number	
  of	
  searches	
  
per	
  user	
  by	
  0.2%	
  to	
  0.6%.	
  	
  	
  
•  Users	
  do	
  fewer	
  searches	
  the	
  longer	
  they	
  are	
  exposed.	
  	
  
•  For	
  longer	
  delays,	
  the	
  loss	
  of	
  searches	
  persists	
  even	
  aYer	
  latency	
  returns	
  to	
  
previous	
  levels.	
  
Performance	
  Measurement	
  
•  Top	
  Percen-le	
  (TP)	
  is	
  a	
  method	
  to	
  find	
  
minimum	
  @me	
  	
  in	
  which	
  (X)	
  amount	
  of	
  
requests	
  were	
  served.	
  
– Example:	
  If	
  the	
  TP90	
  of	
  a	
  page	
  is	
  2	
  sec,	
  then	
  it	
  
means	
  90%	
  of	
  the	
  requests	
  were	
  served	
  within	
  2	
  
seconds.	
  
RUM	
  with	
  SPA	
  
•  What	
  is	
  RUM?	
  
–  RUM	
  =	
  Real	
  User	
  Monitoring,	
  is	
  the	
  actual	
  measure	
  of	
  how	
  long	
  a	
  page	
  
took	
  to	
  render	
  and	
  become	
  usable	
  for	
  a	
  real	
  user.	
  It	
  is	
  the	
  measure	
  of	
  
a	
  page’s	
  speed	
  by	
  measuring	
  the	
  @me	
  it	
  takes	
  for	
  "PAGE	
  READY"	
  state,	
  
and	
  not	
  only	
  the	
  @me	
  it	
  took	
  for	
  data	
  to	
  be	
  downloaded.	
  
•  Web	
  is	
  moving	
  towards	
  Single	
  Page	
  Architecture	
  
•  Most	
  of	
  Exis@ng	
  tools	
  stop	
  at	
  the	
  onLoad()	
  and	
  are	
  not	
  sufficient	
  to	
  get	
  the	
  
real	
  measure	
  of	
  the	
  end	
  user	
  performance	
  and	
  cannot	
  tell	
  the	
  PAGE	
  
READY	
  state	
  
•  SPA	
  world	
  -­‐	
  most	
  of	
  the	
  processing	
  starts	
  when	
  onLoad()	
  ends	
  
•  Need	
  to	
  measure	
  processing	
  @me	
  for	
  various	
  user	
  ac@ons	
  heavy	
  on	
  the	
  
client	
  side	
  
Measured	
  Availability	
  using	
  NewRelic	
  
RUM	
  
9	
  
Web	
  Servers	
  
With	
  NewRelic	
  agents	
  
MA	
  reads	
  metrics	
  from	
  
NewRelic	
  insights	
  
NewRelic	
  Plugin	
  sends	
  
page	
  load	
  9me,	
  url,	
  
success	
  code	
  to	
  NewRelic	
  
servers	
  
Customer	
  interac9ons	
  
with	
  Web	
  servers.	
  
Web	
  servers	
  respond	
  to	
  
customer’s	
  requests	
  
along	
  with	
  NewRelic	
  
RUM	
  plugin.	
  
Measured	
  
Availability	
  
Dashboard	
  
Customer	
  
Browser	
  
Measured	
  Availability	
  Formulae	
  
	
  	
  
	
  
	
  
	
  
10	
  
NOTE:	
   “Outage”	
   is	
   number	
   of	
   transac@ons	
   that	
   were	
   impacted	
   during	
   planned	
   or	
  
unplanned	
  down@me.	
  
MA(%)=
(TotalTransactions−(ServerErrors+SlowTrasanctions+Outage))
TotalTransactions
×100
Thank	
  You!	
  
Singh.sanjay@yahoo.com	
  
+91.96633.72266	
  	
  

More Related Content

PPTX
Building the right website monitoring strategy
PPTX
Optimising your web application
PPTX
Performance analysis in eCommerce (retail and Websphere Commerce)
 
PDF
09 performance
PPT
How does proxy works?
PPTX
Service workers and their role in PWAs
PDF
Introduction core web vitals
PDF
Poslovni Imenik BiH - Lokal d.o.o.
Building the right website monitoring strategy
Optimising your web application
Performance analysis in eCommerce (retail and Websphere Commerce)
 
09 performance
How does proxy works?
Service workers and their role in PWAs
Introduction core web vitals
Poslovni Imenik BiH - Lokal d.o.o.

Viewers also liked (19)

PDF
How to design a great dashboard (Hebrew)
PPTX
New Relic Future Stack 2015 - Step 1 in our quest for DevOps at US Foods 111...
PDF
DevconTLV 2014 (Jan) - DIY DevOps
PDF
Customer Ops: DevOps <3 customer support
PPTX
Survey on article extraction and comment monitoring techniques
PDF
DevOps at Obama for America(2012) and the DNC (DevOps Days NYC Jan 2013)
PDF
Practical Monitoring Techniques
PDF
Which watcher watches CloudWatch
PDF
5 Ways ITSM can Support DevOps, an ITSM Academy Webinar
PDF
DevOps Roadtrip Minneapolis
PDF
DevOps/Flow workshop for agile india 2015
PDF
Devoxx 2014 monitoring
PDF
Run IT Support the DevOps Way
PDF
Jelastic - DevOps PaaS Business with Docker Support for Service Providers
PDF
Paris Devops - Monitoring And Feature Toggle Pattern With JMX
PDF
Devops the Microsoft Way
PDF
DevOps in the Cloud with Microsoft Azure
PPTX
DevOps monitoring: Feedback loops in enterprise environments
PDF
VIZBI 2015 Tutorial: Cytoscape, IPython, Docker, and Reproducible Network Dat...
How to design a great dashboard (Hebrew)
New Relic Future Stack 2015 - Step 1 in our quest for DevOps at US Foods 111...
DevconTLV 2014 (Jan) - DIY DevOps
Customer Ops: DevOps <3 customer support
Survey on article extraction and comment monitoring techniques
DevOps at Obama for America(2012) and the DNC (DevOps Days NYC Jan 2013)
Practical Monitoring Techniques
Which watcher watches CloudWatch
5 Ways ITSM can Support DevOps, an ITSM Academy Webinar
DevOps Roadtrip Minneapolis
DevOps/Flow workshop for agile india 2015
Devoxx 2014 monitoring
Run IT Support the DevOps Way
Jelastic - DevOps PaaS Business with Docker Support for Service Providers
Paris Devops - Monitoring And Feature Toggle Pattern With JMX
Devops the Microsoft Way
DevOps in the Cloud with Microsoft Azure
DevOps monitoring: Feedback loops in enterprise environments
VIZBI 2015 Tutorial: Cytoscape, IPython, Docker, and Reproducible Network Dat...
Ad

Similar to Measured availability - Sanjay Singh - DevOps Bangalore meetup March 28th 2015 (20)

PPTX
A Modern Approach to Performance Monitoring
PDF
Synthetic and RUM - Best of bo
PDF
Walmart pagespeed-slide
PDF
Walmart Web Performance Circa 2013
PDF
Edge 2014: A Modern Approach to Performance Monitoring
PDF
Improving user experience with real user measurements
PDF
The Critical Path to Performance: User Journeys
PPTX
Web performance and measurement - UKCMG Conference 2011 - steve thair
PPTX
Radial | SOASTA IR Webinar
PPTX
Metrics, Metrics Everywhere (but where the heck do you start?)
PPTX
Metrics, Metrics Everywhere (but where the heck do you start?)
PPTX
Metrics, metrics everywhere (but where the heck do you start?)
PPTX
Metrics, metrics everywhere (but where the heck do you start?)
PPTX
Velocity NYC: Metrics, metrics everywhere (but where the heck do you start?)
PPTX
Metrics, metrics everywhere (but where the heck do you start?)
PPTX
Measuring Front-End Performance - What, When and How?
PPTX
Measuring web performance with user-centric metrics
PPTX
PAC 2020 Santorin - Ankur Jain
PPTX
Measuring performance - Velocity 2016 Training
PPT
Watching websites
A Modern Approach to Performance Monitoring
Synthetic and RUM - Best of bo
Walmart pagespeed-slide
Walmart Web Performance Circa 2013
Edge 2014: A Modern Approach to Performance Monitoring
Improving user experience with real user measurements
The Critical Path to Performance: User Journeys
Web performance and measurement - UKCMG Conference 2011 - steve thair
Radial | SOASTA IR Webinar
Metrics, Metrics Everywhere (but where the heck do you start?)
Metrics, Metrics Everywhere (but where the heck do you start?)
Metrics, metrics everywhere (but where the heck do you start?)
Metrics, metrics everywhere (but where the heck do you start?)
Velocity NYC: Metrics, metrics everywhere (but where the heck do you start?)
Metrics, metrics everywhere (but where the heck do you start?)
Measuring Front-End Performance - What, When and How?
Measuring web performance with user-centric metrics
PAC 2020 Santorin - Ankur Jain
Measuring performance - Velocity 2016 Training
Watching websites
Ad

Recently uploaded (20)

PDF
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
PDF
Agricultural_Statistics_at_a_Glance_2022_0.pdf
PPTX
Spectroscopy.pptx food analysis technology
PDF
Network Security Unit 5.pdf for BCA BBA.
PDF
Encapsulation_ Review paper, used for researhc scholars
PPTX
ACSFv1EN-58255 AWS Academy Cloud Security Foundations.pptx
PPTX
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
PDF
Machine learning based COVID-19 study performance prediction
PDF
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
PDF
Chapter 3 Spatial Domain Image Processing.pdf
PDF
Reach Out and Touch Someone: Haptics and Empathic Computing
PDF
Optimiser vos workloads AI/ML sur Amazon EC2 et AWS Graviton
PDF
Unlocking AI with Model Context Protocol (MCP)
PDF
cuic standard and advanced reporting.pdf
PPTX
sap open course for s4hana steps from ECC to s4
PPT
Teaching material agriculture food technology
PPTX
Effective Security Operations Center (SOC) A Modern, Strategic, and Threat-In...
PDF
Empathic Computing: Creating Shared Understanding
PPTX
Digital-Transformation-Roadmap-for-Companies.pptx
PPTX
Programs and apps: productivity, graphics, security and other tools
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
Agricultural_Statistics_at_a_Glance_2022_0.pdf
Spectroscopy.pptx food analysis technology
Network Security Unit 5.pdf for BCA BBA.
Encapsulation_ Review paper, used for researhc scholars
ACSFv1EN-58255 AWS Academy Cloud Security Foundations.pptx
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
Machine learning based COVID-19 study performance prediction
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
Chapter 3 Spatial Domain Image Processing.pdf
Reach Out and Touch Someone: Haptics and Empathic Computing
Optimiser vos workloads AI/ML sur Amazon EC2 et AWS Graviton
Unlocking AI with Model Context Protocol (MCP)
cuic standard and advanced reporting.pdf
sap open course for s4hana steps from ECC to s4
Teaching material agriculture food technology
Effective Security Operations Center (SOC) A Modern, Strategic, and Threat-In...
Empathic Computing: Creating Shared Understanding
Digital-Transformation-Roadmap-for-Companies.pptx
Programs and apps: productivity, graphics, security and other tools

Measured availability - Sanjay Singh - DevOps Bangalore meetup March 28th 2015

  • 1. Measured  Availability  and   Performance       Sanjay  Singh   singh.sanjay@yahoo.com  
  • 2. Agenda   •  FCI  –  Failed  Customer  Interac@on   •  TP  –  Top  Percen@le   •  RUM   •  RUM  with  SPA  
  • 4. FCI   •  FCI  is  Failed  Customer  Interac@on  which  is   measured  for  a  par@cular  page/workflow.   – A  customer  interac@on  can  be  page  view,  form   submission,  etc.  In  HTTP  terms,  customer’s   interac@ons  are  GET/POST  request.   – If  the  page  response  is  not  as  per  the  expected   behavior  of  the  applica@on,  then  it’s  a  failed   customer  interac@on.  
  • 5. How  is  FCI  calculated  ?   •  FCI  is  calculated  on  per  page  basis.  If  page   execu@on  is  not  as  per  the  expected  behavior,   then  it  is  counted  as  failure.   •  Iden@fy  all  failure  cases  and  log  it   appropriately  on  the  server  with  user  session’s   info.  
  • 6. Performance   •  As  per  Google  –   •  Increasing  web  search  latency  100  to  400  ms  reduces  the  daily  number  of  searches   per  user  by  0.2%  to  0.6%.       •  Users  do  fewer  searches  the  longer  they  are  exposed.     •  For  longer  delays,  the  loss  of  searches  persists  even  aYer  latency  returns  to   previous  levels.  
  • 7. Performance  Measurement   •  Top  Percen-le  (TP)  is  a  method  to  find   minimum  @me    in  which  (X)  amount  of   requests  were  served.   – Example:  If  the  TP90  of  a  page  is  2  sec,  then  it   means  90%  of  the  requests  were  served  within  2   seconds.  
  • 8. RUM  with  SPA   •  What  is  RUM?   –  RUM  =  Real  User  Monitoring,  is  the  actual  measure  of  how  long  a  page   took  to  render  and  become  usable  for  a  real  user.  It  is  the  measure  of   a  page’s  speed  by  measuring  the  @me  it  takes  for  "PAGE  READY"  state,   and  not  only  the  @me  it  took  for  data  to  be  downloaded.   •  Web  is  moving  towards  Single  Page  Architecture   •  Most  of  Exis@ng  tools  stop  at  the  onLoad()  and  are  not  sufficient  to  get  the   real  measure  of  the  end  user  performance  and  cannot  tell  the  PAGE   READY  state   •  SPA  world  -­‐  most  of  the  processing  starts  when  onLoad()  ends   •  Need  to  measure  processing  @me  for  various  user  ac@ons  heavy  on  the   client  side  
  • 9. Measured  Availability  using  NewRelic   RUM   9   Web  Servers   With  NewRelic  agents   MA  reads  metrics  from   NewRelic  insights   NewRelic  Plugin  sends   page  load  9me,  url,   success  code  to  NewRelic   servers   Customer  interac9ons   with  Web  servers.   Web  servers  respond  to   customer’s  requests   along  with  NewRelic   RUM  plugin.   Measured   Availability   Dashboard   Customer   Browser  
  • 10. Measured  Availability  Formulae             10   NOTE:   “Outage”   is   number   of   transac@ons   that   were   impacted   during   planned   or   unplanned  down@me.   MA(%)= (TotalTransactions−(ServerErrors+SlowTrasanctions+Outage)) TotalTransactions ×100
  • 11. Thank  You!   Singh.sanjay@yahoo.com   +91.96633.72266