2015 Tableau Server on AWS (Tableau Customer Conference)
- 3. Tableau Server on AWS Checklist for Deployment Jason Frankenfield Data Analytics Systems Architect The Climate Corporation # C l i m a t e C o r p D a t a Albert Wong Reporting Platform Manager Netflix
- 4. SF Bayarea Tableau User Group
- 5. Where is your data? Climate: RDS Postgres, Redshift Netflix: Teradata, S3, RDS MySQL The point: most, if not all datasources are already in cloud if not, VPC links to AWS highly recommended Where are your users? Climate: 150 Tableau users, 5 offices across the US Netflix: 800 Tableau users, mainly US, Europe, Asia Why AWS? Fast deployment Huge technology ecosystem with many services/options Easy to experiment, with caveats
- 6. Viz Tools Tableau Microstrategy Data Sources Presto Hive Teradata Redshift Tableau Server R3.4xlarge 1 gateway, 3 workers Viz Tools Tableau PeriscopeData Data Sources RDS PostGreSQL Redshift Tableau Server R3.2xlarge 1 gateway, 1 worker
- 7. AWS Environment: EC2-VPC vs. EC2 Classic EC2-VPC: private cloud subnet, able to tunnel to your corp network EC2-Classic: flat, public network layout Single Server vs. Multi Server Single Server Simple to set up Doesn’t require AD, good for Core License users Resource contention may become an issue Multi Server More complex - requires AD, security group permissions Allocate roles between servers for best performance
- 8. Tableau Options Active Directory: Easy, but requires users to type password in again SAML Single Signon: - More Complex, requires IDP service Local Authentication: Simple, but more management overhead Climate Corp Active Directory via VPC link to corp network Netflix SAML - Meechum (SSO on PingFederate platform) Don’t have in-house AD? Amazon Directory Service: https://aws.amazon.com/directoryservice/
- 9. Create groups and rules for: Server access to data sources User access to Server Admin access to Server hosts (RDP) Server-Server communication Recommended: Security Group: Server-server access + Admin RDP Security Group: User Access Security Group: Data source access Pitfalls: Cross-account security groups can be tricky Server will use the same DNS names and port as Publisher’s data connection Make sure ALL servers have access to data sources
- 10. Provides reliable DNS name and IP for server access Replace or upgrade your server easily Required for HA or multiple gateway Server configs Upgrade to HA/multiple gateways at any time Pitfall: set gateway.public.host and gateway.public.port config options Protip: use ELB for SSL offload Server upgrade made easy Build new servers ahead of time Take old server(s) out of ELB Take backup Copy to new main server and restore Add new servers(s) to ELB
- 11. Not possible to do 100% right now Licensing Admin user creation Add/config worker servers Reduce your time to restore Document your config to text file with AWS CLI http://docs.aws.amazon.com/cli/latest/reference/ec2/describe-instances.html http://docs.aws.amazon.com/cli/latest/reference/elb/describe-load-balancers.html Use AWS CLI instead of the GUI Create a custom AMI with drivers/installers Upload tabadmin backup files to S3
- 12. Don’t just terminate old server instances Deactivate Tableau license before termination Don’t install on C: Install to non-OS makes it easy to upgrade/resize volumes later with EBS snapshot/restore Human errors happen! Don’t lose your instance Termination Behavior: Shutdown Termination Protection: On Make regular backups Run tabadmin backup regularly Copy backup files to file server or S3
- 13. Monitoring Cloudwatch Set thresholds + alert via email CPU monitoring by default run scripts in host OS for more metrics/alerts http://docs.aws.amazon.com/AmazonCloudWatch/latest/DeveloperGuide/mon-scripts-powershell.html Notification about Service/Application down is critical Backups tabadmin backup + S3 http://docs.aws.amazon.com/powershell/latest/userguide/pstools-s3-upload-object.html Palette Tableau management service
- 14. Upgrade Instance type Add RAM, CPU with a couple clicks Requires instance shutdown/restart Running out of space? Snapshot the volume Restore to a larger volume Expand partition via Windows Disk Management Experiment and tweak Experiment with Server process allocation Tune your data sources http://d0.awsstatic.com/whitepapers/redshift/Tableau_Redshift_Whitepaper.pdf
- 15. Make sure AWS is right for you Data Sources Users User Management Use the management system available to you Security Groups are critical access to data sources and users is key to success Use an Elastic Load Balancer You can’t automate end-to-end But you can shorten deploy/restore timeTune your data sources Monitor and Backup your server Experiment!
- 16. Please complete the session survey from the Session Details screen in your TC15 app