Monitoring Solutions for APIs
Download as PPTX, PDF5 likes916 views
How to monitor and manage the API experience with Apigee's tools to measure API latencies, manage SLAs, and more. At I Love APIs 2015
Ad
More Related Content
What's hot (20)
Similar to Monitoring Solutions for APIs (20)
Ad
More from Apigee | Google Cloud (20)
Ad
Recently uploaded (20)
Monitoring Solutions for APIs
- 1. 1 Monitoring Solutions for APIs Arvind Jain, Apigee Vinoth Pethaiyan, Apigee
- 2. Build your APIs, and they will come … hopefully
- 3. First come the Developers to build apps using your APIs 3©2015 Apigee. All Rights Reserved.
- 4. Then the users … who use all those apps and your APIs 4©2015 Apigee. All Rights Reserved.
- 5. • Are the APIs up? • How are they performing? • How’s the user experience? • How do I measure errors? • Can I make any guarantees? That’s not enough … watch over your APIs to keep users happy 5©2015 Apigee. All Rights Reserved.
- 6. What failed? Who’s responsible? How to resolve? Stuff happens!!! What to do when something fails? • Which APIs? • Whose apps? • How widespread? • App teams • API developers • Vendors … open a P1 • Backend system teams 6 • Control traffic • Restart systems • Modify code ©2015 Apigee. All Rights Reserved.
- 7. R R MP MP C* C* Region N R R MP MP C* C* Region 1 Anatomy of an API call 7 Apigee infrastructure Incoming app traffic API proxies and policies Backend targets Valid requests 200 Invalid syntax Wrong password Expired token Bad client requests 4xx Proxy code Policy violation Callout errors Failure 5xx Memory Bugs Server down Failure 5xx Long latency Timeouts Server down Failure 5xx
- 8. Part of the API lifecycle ©2015 Apigee. All Rights Reserved. 8 Design Develop Secure PublishMonitor Analyze Monetize Scale
- 9. Monitor and manage the API experience 9©2015 Apigee. All Rights Reserved. Devices & apps API Backend Measure API latencies Manage API SLAs Fast response time Correct responses Global consistency Guarantee success rate Manage risk Make commitments
- 11. Users Developers Why does it matter? 61% expect apps to start in < 4 sec. 49% expect apps to respond in < 2 sec. 80% will retry problem apps up to 3 times > 40% apps depend on 5+ cloud services 11©2015 Apigee. All Rights Reserved. Source: 2015 Mobile App Usage Abandonment Survey Source: Mobile Experience Benchmark Report, Crittercism
- 12. US 1.0x Africa 1.9x Canada 0.8x Europe 1.2x Middle-East 1.5x China 1.1x Japan 0.9x Australia 1.2x SE Asia 1.6x India 2.1x >2x global variation in response times 12 Source: Mobile Experience Benchmark Report, Crittercism Mobile app response times
- 13. How do you measure latencies globally? 13 Edge of Network Synthetic API calls Washington DC - US Oregon - US Dublin - Ireland Sydney - Australia APIs 1.Measure response times 2.Validate response codes 3.Compare across regions 4.Alerts upon problems API Health* Quick Check-Up for your APIs * Beta
- 14. Improve your API proxies • Are your APIs and policies optimized? • Do you really need that extra policy? Deploy APIs across multiple regions • What’s the global distribution of your users? • Is your API infrastructure optimized to deliver the best experience for all users? What can you do about it? 14©2015 Apigee. All Rights Reserved.
- 15. Single region deployment 15 Great if all users are local Large variance in user experience
- 16. Multi-region deployment 16 Consistent user experience
- 17. Demo – API Health 17©2015 Apigee. All Rights Reserved.
- 18. Managing SLAs 18
- 19. Affected APIs Impacted developers & apps Source of errors How do you guarantee something you don’t know… • Individual APIs or • Entire deployment • Registered developers • Associated apps 19 • API policies • Runtime infrastructure • Backend systems ©2015 Apigee. All Rights Reserved.
- 20. Measure Investigate Mitigate Meant for DevOps and API producers • API availability and establish SLAs with customers • Real-time visibility into API traffic & problems • Isolate API problems • Intelligent alerting 20 • Apply restoration strategies • Mitigate impact of API problems ©2015 Apigee. All Rights Reserved.
- 21. Capture responses for all API traffic – Very important! – Categorize by APIs, developers, and apps Separate client errors from server errors – Client errors do not affect API availability Measure API availability 21©2015 Apigee. All Rights Reserved. Distinguish server errors across all tiers – API policies – Runtime infrastructure – Backend systems Continuously calculate for current view – Every 5 mins – Averaged over longer periods of time
- 22. Current-time view of traffic metrics – Collect at source – HTTP status codes – Latencies – Resource usage – Events – Configuration Identify problems and dependencies – Volume of traffic – Availability misses – Errors – Latencies – Regions – Backend systems Investigate 22©2015 Apigee. All Rights Reserved. Compare across multiple tiers – Instances – Tiers • APIs • Infrastructur e • Backend targets Diagnose individual transactions – Activate real- time trace – Investigate step-by-step latencies
- 23. Mitigate • Rapid response to API traffic issues – Administered by Ops teams without requiring proxy editing – Remediation actions to temporarily restore API traffic • Isolate offending parties – App blocking – respond to rogue or compromised apps or keys – Traffic rerouting – redirect inbound traffic to new location • Graceful degradation – Traffic throttling – reduce flow to backend systems – Fail fast – reduce long time-outs
- 24. Thank you 24©2015 Apigee. All Rights Reserved.
Editor's Notes
- #3: Hope… Pray… Plan
- #5: Are you prepared?
- #10: Traditional view of Availability based on uptime is flawed System can be up but API calls could be really slow or failing SLA = guarantee that API Success Rate > x% API Success Rate = % API calls successfully handled Manage risk and make commitments API availability is critical to user experience Unexpected API errors == potential lost revenues Single metric measures quality of end-2-end API delivery network
- #12: http://www.slideshare.net/TechBeacon/2015-mobile-app-usage-and-abandonment-survey http://pages.crittercism.com/mobile-app-experience-benchmark-report-thank-you.html?aliId=31677190
- #14: API Health has a set of probes that call into your APIs from multiple data centers around the world to measure their response time and responses