MySQL and MongoDB Monitoring with PMM

1 . MySQL and MongoDB Monitoring and Optimization with Percona Monitoring and Management Peter Zaitsev CEO Percona University, Ghent June 22nd, 2017 1 © 2017 Percona 

2 .PMM Stands for 2 © 2017 Percona 

3 .Why Did We Start Working on PMM 3 © 2017 Percona 

4 .Percona Vision 100% Free and Open Source Platform for 95% of applications 4 © 2017 Percona 

5 .Monitoring and Management Critical need for modern database infrastructure 5 © 2017 Percona 

6 .Monitoring and Management Status Commercial License Cloud Based Do it yourself 6 © 2017 Percona 

7 .Would be great to have 100% Free and Open Source Can run In the Cloud and on Premises Easy to use 7 © 2017 Percona 

8 .Available for Anyone on the Any Environment Team •Engineers, •Development •DBA •Staging/QA •SREs •Production 8 © 2017 Percona 

9 .About PMM 9 © 2017 Percona 

10 .PMM Philosophy Do not reinvent the wheel Use as much of industry leading components as possible Make Integration and Customization easy 10 © 2017 Percona 

11 .Industry Leading Components 11 © 2017 Percona 

12 .Current Focus To Become Best 100% Free and Open Source Solution for MySQL and MongoDB 12 

13 .Currently Works Metrics for MySQL and MongoDB Support for PXC, Galera, ProxySQL Support of Amazon RDS Query Analytics for MySQL Experimental Orchestrator Integration Do it yourself Grafana Alerting 13 © 2017 Percona 

14 .PMM Platform Support Server • Docker • Virtual Appliance Side • AMI Client Side • Linux 14 © 2017 Percona 

15 .In Works Support for Longer retention period Better Cloud Support Ease of use Query Analytics for MongoDB Improved Query Analytics Interface Alerting 15 © 2017 Percona 

16 .By the Experts for the Experts ? Assist Experts to Resolve the most complicated problems But be friendly enough for Non-Expert use Actionable Advice Integration and Automation 16 © 2017 Percona 

17 .PMM Data Capture 1 sec resolution data capture for most important data Capture a lot more with lower resolution Allow to perform deep analyses 1000s of metrics captured per instance 17 © 2017 Percona 

18 .PMM Architecture Overview 18 © 2017 Percona 

19 .Database as a Black Box…. 19 © 2017 Percona 

20 .What does MySQL and MongoDB do ? Responds Application Requests 20 © 2017 Percona 

21 .Criteria Response (Available) Right Quickly (Correctness) (Performance) 21 © 2017 Percona 

22 .Problem Caused By Hardware and Application Database Environment • Too many • Choosing Bad • CPU Queries Plan • Disk • Bad Queries • Contention • Memory • Improper Issues • Network Capacity • Locking Planning 22 © 2017 Percona 

23 .Query Analyses Working for MySQL Will be available for MongoDB shortly 23 © 2017 Percona 

24 .What Queries are causing the load ? 24 © 2017 Percona 

25 .Why are they causing this load ? 25 © 2017 Percona 

26 .How to fix them 26 © 2017 Percona 

27 .System Information 27 © 2017 Percona 

28 .What happens on OS and Hardware Level 28 © 2017 Percona 

29 .As well as Database Level 29