Skip to content

Latest commit

 

History

History
86 lines (53 loc) · 2.9 KB

04-Logging-Scaling-High-Availability.md

File metadata and controls

86 lines (53 loc) · 2.9 KB
Raw

Logging, Scaling, and High Availability

  • How do you access application logs?

cf logs APP_NAME

  • What are the components of the Loggregator system?
  1. Metron - Forward logs to the Loggregator subsytem
  2. Doppler - gather logs from Metron
  3. Traffic Controller - handles client requests for logs
  4. Firehose - a websocket endpoint that exposes app logs, container metrics and ER component metrics
  5. Nozzles - consume the firehose output
  • How do you scale an application manually?

cf scale <APP_NAME> -m -i <Instance_COUNT>

  • What are the four levels of high-availability provided by PCF?
  1. BOSH detects and recreates failed VMs
  2. BOSH also detects and restarts failed processes
  3. Application instances are automatically spread -- Across different Cells -- Across Cells in different Availability Zones
  4. CF Converger detects if an application instance has failed and restarts it
  • What is the difference between scaling up and scaling out?
  • scale up (vertically) increases the disk space limit or memory limit of all instances of an application. It requires downtime as the container is recreated.
  • scale out (horizontally) increases the number of instances of an application by creating or destroying them. It requires no downtime.

pluralsight recap

  • What are some fundamental changes impacting application design and delivery?
  • Distributed computing
  • Ephemeral infrastructure
  • Immutable infrastructure
  • What is a modern methodology for delivering cloud based applications?

Twelve-factor app

  • How are logs handled in a Twelve-factor app?

As event streams

  • True or false, Garden is an interface with many implementations.

True, there's Linux and Windows and more on the way

  • Which component of the Elastic runtime exposes an API for users to manage applications?

Cloud controller

  • What are the four levels of high availability that keep your applications running?
  • Availability zones
  • BOSH managed processes
  • Failed VMs - any VM that BOSH deploys will be recreated if you enable resurrector
  • Self healing application instances handled by the brain within Diego
  • Where should you write your application logs?
  • To Standard Out and Standard Error
  • What are some of the different origin codes that were seen in the log?
  • cf logs
  • API for cloud controller
  • stage for stating
  • cell for running applications

  • What did you think about accessing the logs? How is that different from what you do today?

  • What is the difference between scaling out versus scaling up?

Scaling up requires down time to recreate the container Scaling horizontally requires no downtime

  • How do you recover failed application instances today? What happens if the underlying platform fails? How do you recover from that?

  • How could you tell if your application has been crashing?

Look at the logs, but probably the best place to go is cf events