As an FME Server Administrator, part of your role is to oversee the efficient processing of jobs. But as your FME Server usage grows, you may face instances when the demand exceeds your engine processing capacity and jobs begin to pile up in the queue. If you have ever wondered:
How can I make my jobs more performant?
Why is my job taking longer to run on FME Desktop than FME Server?
How can I prioritize short over long jobs or those submitted by certain users?
How can I prevent engines from sitting idle while I have jobs waiting in the queue?
Then this presentation is the one for you. We’ll highlight different tools you can use to monitor both the individual job performance and the overall job throughput. We’ll also show you what methods you can implement for improved resource and job optimization.
2. Meet the Presenters
Holly Coxon
Technical Support Team Lead, FME
Server
Merline George
Technical Support Specialist, FME
Server
3. Welcome to Livestorm!
A few ways to engage with us during the webinar:
Audio issues? Click this for 4 simple
troubleshooting steps.
*Weʼll also bring up a poll on your
screen during the webinar.
4. How to download slides
1. Hover over the slide deck
in the webinar room
2. Click this button
5. Agenda
● What is Optimization?
● Monitoring FME Server Activity
● Job Performance
● Scaling FME Server Processing Capacity
● Job Prioritization
○ Demo
● Resources
● Q+A
7. Save time and money, and accomplish
more with your FME Engines.
8. What is Optimization?
Optimizing FME Server is an iterative process and requires continuously analysing your
existing setup. Some of the deciding factors for system improvement are:
● job processing time
● system resource usage
● number of engines available
● business requirements
10. When should you employ
optimization techniques?
Businesses need to analyze their FME Server activity in order to make
informed decisions about their system.
11. Monitoring FME Server Activity:
Job Statistics
Job Statistics in FME Server contain information that can be used to better
understand FME Server activity and how Jobs are performing.
● Total Runs
● Average Elapsed Time
● Average % CPU
● Average CPU Time
● Average Peak Memory Usage
Learn more:
Monitoring FME Server Job Activity
using the REST API
2021.0+
12. Monitoring FME Server Activity:
Server Analytics
See historical data on job
throughput and uncover
trends that allow you to
know when and where to
add additional data
processing resources.
*New in 2022.2 Server
Analytics Dashboards for Job
Throughput broken down
per Engine and Queue*
2022.1+
13. Monitoring FME Server Activity:
Customized Dashboards
Learn more:
Monitoring FME Server Job Activity
using the REST API
All Versions
14. What if throughput exceeds capacity?
Businesses need to modify workflows to increase performance and
overall resource optimization.
15. Improvements to workspaces in FME Desktop can increase performance, ultimately freeing up
Engines to process more jobs:
● Small workspace changes with Feature Caching and Feature Counts enabled
● Filter in required attributes as soon as possible
● Avoid grouping memory consuming transformers when dealing with larger datasets
● Avoid chaining duplicate transformers to carry out a different step of one overall calculation
● If you have multiple writers in a workspace, promote the writer handling most data to the front
● Consider partitioning the data into groups and process each separately.
Well Begun is… Half Done
Learn more:
Performance Tuning FME: Workspace
Authoring & Design Patterns
All Versions
16. Breaking Down Jobs to Run Parallel
Learn more:
Level Up With Parallelism in FME
Server
All Versions
17. What is under the Hood?
● Most FME Server components have smaller footprints- 2 GB of RAM and 1
or 2 CPU cores for Core and Application Server component
● Storage required to install FME Server, third party applications and system
processes
Learn more:
FME 2022 Technical Specifications |
Safe Software
What are the hardware requirements for
FME Server?
All Versions
18. ● We recommend ~ 1 CPU core per FME Engine
● Review FME Desktop Workspace logs or Job Statistics on the
Completed Jobs webpage for Engine CPU and memory requirements
Engine Requirements
Learn more: FME Server - Host
System Sizing
19. ● Use SSD storage for local cache/temp storage
● Set the FME_TEMP and FME_HOME paths to local drive.
● Store Workspace Datasets close to FME Engine host or setup reliable fast
network connection between Engine host and data.
Tips & Tricks
Learn more: Setting a temporary
file location for FME to use via
the FME_TEMP environment
variable
FME Server Resources -
Uploading, Referencing and
Configuration
All Versions
20. What if throughput STILL
exceeds capacity?
Businesses need a solution where they can deploy
processing efficiently while only paying for what is used.
21. It’s Time to Scale Up
● FME Server sold by the CPU hour
● Ideal for varied workloads, streaming,
and scalable deployments
CPU-Usage (Dynamic) Engines
● Fixed FME Server processing power
● Ideal for predictable and consistent
workloads
Standard Engines
22. CPU-Usage Pricing
Purchase “CPU time” rather than processing throughput (number of engines) provides:
● Scalability
● Resource Utilization
● Reduce Costs
● Improve Customer Experience
● Agility
2020.0+
Learn more:
Getting Started with CPU-Usage
(Dynamic) Engines
23. How can you meet these varying
processing needs?
Businesses need a solution where they can prioritize jobs and
re-allocate engines based on throughput.
24. The Challenge with Engine Optimization
Jobs are submitted by
automated processes and
on demand by end-users
resulting in a system load
that is highly variable and
changes over time.
25. Three Steps to Job Prioritization
Queue
● Central to FME Server control
● Set Priority
● First in First Out (FIFO)
● Default Queue serves jobs not
assigned to another queue
Job Routing
● Route each job to exactly 1 queue
● Queue based on job properties
including historical information
● Jobs routed based on the first rule
met
Engine Assignment
● Assign engines to 1 or more
queues
● Queue can be assigned based
on engine and queue properties
2021.0+
26. Use Cases
● Additional processing resources for big data lifts
● Stand by capacity for varying processing
demand
● Partition FME Server by department
● Ensure minimal wait times to the end user with
self-serve applications
● Optimize hardware and licensing resources
● Reserved processing resources for data
streaming services
28. ● Optimizing FME Server is a never ending task as new translations are put into
production and your business and processing needs change.
● Analysis of your FME Server activity is key for optimization, and this practice can help
save your business time and money.
● Workspace Authors should be designing workspaces with FME Server in mind to
ensure they are following performant practices.
● FME Server Administrators are responsible for ensuring adequate hardware
provisions and job prioritization configurations.
Thatʼs a Wrap
29. Analyse your
FME Server Job
Activity and
Engine
Utilization
Apply job
performance
strategies and
engine optimization
techniques
Contact us. We
would love to chat
with you about
anything FME
Server
info@safe.com
Next Steps
30. Resources
● Monitoring FME Server Job Activity using the REST API
● FME 2022 Technical Specifications | Safe Software
● FME Server - Host System Sizing
● What are the hardware requirements for FME Server?
● Setting a temporary file location for FME to use via the FME_TEMP environment variable
● FME Server Resources - Uploading, Referencing and Configuration
● Performance Tuning FME
● Performance Tuning FME: Workspace Authoring & Design Patterns
● Level Up With Parallelism in FME Server
● Getting Started with CPU-Usage (Dynamic) Engines
31. Claim Your Community Badge
Get community badges for
watching webinars!
fme.ly/WebinarBadge Todayʼs Code: CFLGB
32. Check out our upcoming
& on-demand webinars:
safe.com/webinars
34. Thank you!
Download FME 2022.1 Free Trial | Upgrade
Chat with us info@safe.com
Connect with us in the Community
Connect with us for more FME
Please share
your feedback
with us through
the webinar
survey!