Since our last product update, we’ve made many improvements to our automated tuning service for PostgreSQL and MySQL. In this latest update, we introduce new features to make sure your database is running correctly and provide insights about how applications are using them. We also added the ability for you to have more control over its configuration tuning process, and make it easier to integrate OtterTune in your production environment. We now discuss these new key features.
Database Health and Insights
The OtterTune auto-tuning engine continuously monitors your database metrics and runs machine learning algorithms to recommend optimization improvements. In this new release, we’ve made OtterTune’s analysis of your database workload more transparent. This transparency comes via three new helpful features:
- Database Health Checks – OtterTune informs you of potentially problematic database behavior.
- Database Insights – OtterTune shows you how it categorizes your database workload, so you can rationalize OtterTune change recommendations, and also make better scheduling and capacity planning decisions.
- Best Configuration Baseline – OtterTune now makes it clear which configuration among its different configuration optimization iterations yielded the best results for your chosen tuning objective.
This release is the first step towards expanding OtterTune to provide you with peace of mind about all facets of your databases.
OtterTune’s Database Health Checks feature automatically provides real-time analysis of your database’s health conditions and suggests actions to fix the issues to improve your database’s performance.
The Database Health checks are unique to different target databases (e.g., PostgreSQL vs. MySQL, and Amazon RDS vs. Amazon Aurora). They inform you about healthy or unhealthy resource utilization, cache, buffer, and I/O efficiency, and other conditions. If your database fails a health check, OtterTune will recommend fixes to the issue.
OtterTune Database Insights reveals how the service categorizes your database workload. It reports metrics that provide you with an overview of your database’s behavior:
- Workload Classification (read-only, read-heavy, write-heavy, write-only, balanced);
- Peak time based on throughput; and
- Disk read and write IOPs.
OtterTune uses this information in its tuning algorithm and health checks. The Insights summary helps you understand why OtterTune makes the recommendations it gives you. If you don’t understand why OtterTune is recommending something, take a look at the panel. The insights can also help you make better maintenance scheduling decisions and help with capacity planning (something OtterTune is working on automating, stay tuned!).
Best Configuration Baseline
OtterTune now automatically tracks the Best Configuration that it has discovered for your database. This is the best configuration OtterTune has seen so far and the latest performance and status information for the database. To help you maximize database performance, this panel includes statistics on measured target objectives for the database’s best configuration, percentage improvement of target objective value of the best configuration over the baseline (first) configuration, and the name and status of the best configuration.
To enable this feature:
- The tuning interval must be at least 24 hours.
- The baseline (first) configuration has at least 24 hours of target objective data; and,
- The target objective cannot be throughput.
New Target Objective: CPU Utilization
We also support more Target Objectives for OtterTune’s tuning algorithm: p99 query latency (the recently announced default) and all-new CPU utilization. The Target Objective is what OtterTune optimizes for when it generates new configurations for the database. You can change the Target Objective to guide OtterTune towards optimizing performance versus cost, for example.
Auto-Tune Aurora for Free
Amazon Aurora support is now available with our free Starter tier plan, which offers basic auto-tuning of the most common knobs. We now support Aurora MySQL v5.6 and v5.7 (v8.0 coming soon!) and Aurora PostgreSQL versions v10, v11, v12, and v13.
You can get started auto-tuning Aurora for free by creating an OtterTune account and following the quickstart tutorial. Please refer to the OtterTune pricing information for auto-tuning more than one database or additional enterprise features.
Auto-Tune PostgreSQL v14
Those of you running the latest version of PostgreSQL can now use OtterTune to auto-tune your databases and monitor database health. We’ve also added support for new configuration knobs introduced in PostgreSQL 14. Here’s the list of 100+ PostgreSQL knobs that OtterTune optimizes.
More Integration Options
We’ve also added more integration options to make OtterTune a seamless part of your environment. We offer Kubernetes support for deploying the OtterTune Agent, so you can connect to your database privately and securely, and send its knob/metric data back to the OtterTune service.
We’re also streamlining identity and access management with Terraform support for IAM role connectivity in addition to support for CloudFormation. Setup for Terraform can either be done manually or automatically.
User Docs and an Open Support Community on Slack
Check out our OtterTune user documentation, which details how to connect OtterTune to your database and start the machine learning (ML)-driven tuning process.
Create your free OtterTune account and join us on Slack to let us know how it’s going. We look forward to your questions and feedback!