avoid the fail whale - design for availability

50
Brian O’Connell @boc_tothefuture linkedin.com/in/broconne Avoid the Fail Whale Design for Availability

Upload: brian-oconnell

Post on 06-Apr-2017

78 views

Category:

Engineering


3 download

TRANSCRIPT

Brian O’Connell@boc_tothefuture

linkedin.com/in/broconne

Avoid the Fail Whale

Design for Availability

2x

Cyber Monday

+

Black Friday

$17.8B / day

2x

Cyber Monday

+

Black Friday

$972k / sec

2x

Cyber Monday

+

Black Friday

$17.8B / day

Domain Cloud Operations Application

Sports

Enterprise

Domain Cloud Operations Application

Hybrid

Cloud

Domain Cloud Operations Application

Cloud

+

Ops+

Dev

Domain Cloud Operations Application

3 or 4 Locations

Domain Cloud Operations Application

$$$

2 > 3

Domain Cloud Operations Application

Identical

Domain Cloud Operations Application

Automate

Domain Cloud Operations Application

No HA Takeover

Domain Cloud Operations Application

Global Traffic

Management

Domain Cloud Operations Application

Stand Ups

Domain Cloud Operations Application

Global

Staff

Domain Cloud Operations Application

Technical

Leader

Domain Cloud Operations Application

Squads

Domain Cloud Operations Application

Proactive

Monitoring

Domain Cloud Operations Application

Change

Orchestration

Domain Cloud Operations Application

Business Hours

Changes

Domain Cloud Operations Application

Slack

IRC

Domain Cloud Operations Application

Single Pane

of Glass

Domain Cloud Operations Application

De-Advertise

Domain Cloud Operations Application

Maintenance

Domain Cloud Operations Application

CAP Theorem

Domain Cloud Operations Application

Consistency

Domain Cloud Operations Application

Availability

Domain Cloud Operations Application

Partition

Tolerance

Domain Cloud Operations Application

Systems of

Engagement

Domain Cloud Operations Application

Financial Data

Domain Cloud Operations Application

Session

Replication

Domain Cloud Operations Application

Bi – Directional

P2P Replication

Domain Cloud Operations Application

NoSQL

Domain Cloud Operations Application

Object Storage

Domain Cloud Operations Application

Non-Destructive

Updates

Domain Cloud Operations Application

Embrace

Eventual

Consistency

Domain Cloud Operations Application

Dev/Ops

Domain Cloud Operations Application

Continuous

Integration

Domain Cloud Operations Application

Continuous

Delivery

Domain Cloud Operations Application

Incremental

Rollouts

Domain Cloud Operations Application

Design

for

Failure

Domain Cloud Operations Application

Share

Nothing

Domain Cloud Operations Application

Logging

Domain Cloud Operations Application

Service

Discovery

Domain Cloud Operations Application

Ephemeral

Resources

Domain Cloud Operations Application

Avoid OS

Lock in

Domain Cloud Operations Application

Blameless

Post Mortem

Domain Cloud Operations Application

Google these:“Always On: Assess, Design, Implements and Manage Continuous Availability”, Herbie Pearthree“Top 9 rules for Cloud Applications”, Kyle Brown

Deep Dive

20

IBM’s statements regarding its plans, directions, and intent are subject to change or withdrawal without notice and at IBM’s sole discretion.

Information regarding potential future products is intended to outline our general product direction and it should not be relied on in making a purchasing decision. The information mentioned regarding potential future products is not a commitment, promise, or legal obligation to

deliver any material, code or functionality.

Information about potential future products may not be incorporated into any contract.

The development, release, and timing of any future features or functionality described for our products remains at our sole discretion.

Performance is based on measurements and projections using standard IBM benchmarks in a controlled environment. The actual throughput or performance that any user will experience will vary depending upon many factors, including considerations such as the amount of

multiprogramming in the user’s job stream, the I/O configuration,

the storage configuration, and the workload processed. Therefore, no assurance can be given that an individual user will achieve results similar to those stated here.

Brian O’Connell@boc_tothefuture

linkedin.com/in/broconne