Are you puzzled about why your server or database performance has suddenly degraded or become slow? This can be quite frustrating, especially when you are trying to maintain optimal performance for your applications
Are you puzzled about why your server or database performance has suddenly degraded or become slow? This can be quite frustrating, especially when you are trying to maintain optimal performance for your applications. There could be numerous reasons behind this issue. It is essential to identify the root cause to implement the right solution and get your server back to its peak performance. Let us dive deeper into this and explore potential solutions together.
In this blog post, I will take you on a journey through the comprehensive Azure Advisor performance recommendations specifically tailored for Azure Database for PostgreSQL flexible server. These insights will not only alert you to potential root causes but also guide you in resolving the server issues you are facing. We have you covered from hardware optimizations and logging enhancements to workload management, autovacuum tuning, bloat reduction, and managing inactive replication slots. Let's dive in and unlock the full potential of your server using the performance recommendations provided here.
To ensure you receive proactive notifications regarding Azure Advisor recommendations at the server or subscription level, you have the option to configure alerts to be received via email or text message.
Setting up Alerts for Recommendations
Stay ahead of the game by setting alerts for Azure Advisor recommendations for your Azure Database for PostgreSQL flexible server! Whenever a recommendation is made for one of your resources, an event is triggered, ensuring you are always in the loop. You can customize these alerts to be recommendation-specific, targeting types of recommendations, or set them up at the Azure subscription level to receive notifications for all recommendations related to the resources within the specified subscription. This way, you will never miss a critical insight and can take proactive steps to maintain optimal server performance, ensuring you are always informed and ready to act!
You can set up alerts based on the Impact level, Category, and Recommendation type.
- On Azure portal, search for Advisor and proceed.
- Go to Monitoring and Alerts and click on New Advisor Alert.
- Create the Advisor Alert and set up an Action Group for the recommendation event.
When a recommendation alert is triggered, you will receive notification via email or text message according to your Action Group settings. This alert is sent once per recommendation to ensure that you are informed promptly for timely action.
You have taken the proactive step of setting up an alert to notify you via email if your server enters any vulnerable state. This ensures that you are immediately informed of any potential issues. For instance, you recently received the following recommendation alert in your email, highlighting a specific area that requires attention. This way, you can promptly address any concerns and maintain the optimal performance of your server.
Scenarios in which Azure Advisor recommendation will help keep your Azure Database for PostgreSQL flexible server performant.
Let us explore the various scenarios in which Azure Advisor can proactively assist you in identifying and resolving performance issues with your PostgreSQL server.
Scenario 1- Inactive Logical Replication slots
You have received an alert indicating that your server has an inactive logical replication slot. The outlined section below highlights that this is a high-impact performance. This alert tells me a recommendation named " Review your server for inactive Logical Replication Slots" has been triggered.
Th next step is to go to your server on Azure Portal. You will be able to see Recommendations on the Overview page of your server on the portal.
Click on the highlighted recommendation above. This action will direct you to a detailed page about the recommendation alert you received. On this page, you will find comprehensive information about the recommendation, including the specific details and the reason for receiving it, which is provided in the "Reason" column. The "Reason" column for this recommendation states: "Found inactive logical replication slot(s). Here are the replication slots that have been inactive for the longest: ['myslot']."
Click on the "Recommended Action" to proceed with the necessary steps to resolve your server's vulnerable state.
You will be directed to a page that provides the resolution for the server issue encountered. The necessary action is outlined in the "Recommended Action" section. In this scenario, you will need to drop the inactive logical replication slot as mentioned below.
You can then check your replication slot for it to be active or inactive.
Resolution:
Proceed to remove the inactive logical replication slot as shown below to maintain optimal server performance.
After dropping the inactive replication slot, you will notice within 24 hours that the recommendation no longer appears on the Overview page under Recommendations in your Azure Portal.
Scenario 2 – Checkpoints happening frequently
You have now received an alert indicating that your server has a recommendation stating the server has frequent checkpoint warnings. The outlined section below highlights that this is a high-impact performance. This alert tells me a recommendation named "PostgreSqlCheckpointHappeningFrequently" has been triggered.
Click on the highlighted recommendation above. This action will direct you to a detailed page about the recommendation alert you received. On this page, you will find comprehensive information in the Recommendation details section and the reason for receiving it, which is provided in the 'Reason' column. The 'Reason' column for this recommendation states: 'Checkpoints are occurring frequently over 3 consecutive hours'.
Click on the 'Recommended Action' to proceed with the necessary steps to resolve your server's vulnerable state. The necessary action is outlined in the 'Recommended Action' section. In this scenario, you will need to update your max_wal_size server parameter. Clicking on the 'Recommended Action' will take you to the Server parameters page of your server as shown below.
Search for the max_wal_size parameter and increase it accordingly.
After updating the max_wal_size parameter and resolving the frequent checkpoint warnings, the recommendation will disappear from the Overview page under Recommendations in your Azure Portal within 24 hours.
Scenario 3 – High Bloat ratio
Let us discuss the alert you received regarding a high bloat ratio on your server(s). I will guide you through the troubleshooting process to understand what is happening on the server and how to resolve the issue.
You have now received an alert indicating that your server has a recommendation stating that your server is having a high bloat ratio.
Click on the highlighted recommendation above. This action will direct you to a detailed page about the recommendation alert you received. On this page, you will find comprehensive information in the Recommendation details section and the reason for receiving it, which is provided in the 'Reason' column. The 'Reason' column for this recommendation states: “Here are the Database(s) with the highest BloatRatio: ["testdb"] has a Bloat Ratio(dead/live + dead)[86.08]”.
Click on the 'Recommended Action' to proceed with the necessary steps to resolve your server's vulnerable state. The necessary action is outlined in the 'Recommended Action' section. In this scenario, the page will redirect you to the Troubleshooting guides of your server on the Azure portal. Navigate to the Autovacuum Monitoring tab, where you will find information on Bloat Ratio, Live vs. Dead tuples, Vacuum and Analyze, Autovacuum workers in progress, Autovacuum per table, and Enhanced metrics by server and by database. Navigate through each section to learn more about the server’s bloat and autovacuum information as shown below.
The Autovacuum monitoring troubleshooting guide is designed to assess the health of autovacuum or vacuum on Azure Database for PostgreSQL Flexible Server. During autovacuum execution, it removes dead tuples, freeing up disk space and improving query performance. Additionally, it updates the visibility map to track pages without outdated or deleted data, reducing unnecessary I/O operations.
Another critical function of autovacuum is to prevent transaction ID wraparound situations, which could result in data loss.
By analyzing the results of this troubleshooting guide, you can take proactive steps to optimize autovacuum and maintain the long-term health of your database. As a short-term solution, you can VACUUM ANALYZE the database or specific frequently used tables with high bloat to clean up dead tuples and bring your server to a performant state.
Scenario 4 – Autovacuum turned OFF
Let us discuss the alert you received regarding autovacuum server parameter on your server(s). I will guide you through the troubleshooting process to understand what is happening on the server and how to resolve the issue.
You have now received an alert indicating that your server has a recommendation stating that your server has autovacuum set to OFF.
Click on the highlighted recommendation above. This action will direct you to a detailed page about the recommendation alert you received. On this page, you will find comprehensive information in the Recommendation details section and the reason for receiving it, which is provided in the 'Reason' column. The 'Reason' column for this recommendation states: “atuovacumm is turned OFF, and the Bloat Ratio on these Database(s) ["postgres","testdb"] and Schema(s) ["pg_toast","public"] is greater than 50%”.
Click on the 'Recommended Action' to proceed with the necessary steps to resolve your server's vulnerable state. The necessary action is outlined in the 'Recommended Action' section. In this scenario, you will need to turn on autovacuum server parameter. Clicking on the 'Recommended Action' will take you to the Server parameters page of your server as shown below.
Search for the autovacuum parameter and turn it ON.
Scenario 5 – Wraparound
Let us walk through a scenario where you receive a Wraparound recommendation alert. We will cover how to troubleshoot and resolve the issue to maintain your server's performance.
Click on the highlighted recommendation above. This action will direct you to a detailed page about the recommendation alert you received. On this page, you will find comprehensive information in the Recommendation details section and the reason for receiving it, which is provided in the 'Reason' column. The 'Reason' column for this recommendation states: “The server has utilized 99.85% of transactions. Now 2999999 transactions remaining.”.
Click on the 'Recommended Action' to proceed with the necessary steps to resolve your server's vulnerable state. The necessary action is outlined in the 'Recommended Action' section. In this scenario, you will need to perform the mitigation steps provided in the troubleshooting guide. Clicking on the 'Recommended Action' will take you to the troubleshooting guides. Please visit the Autovacuum Blockers section to get into more details on your server behavior pertaining to wrap around.
Scroll to the bottom of the page to find the Recommendations to resolve the issue as shown in the screenshot below.
Follow through the Recommendations to proactively enable your server to perform effectively.
Scenario 6 – logging parameters
Too much logging always hinders the server’s performance; hence it is important to understand the need for logging and act accordingly.
Let's walk through a scenario where you receive logging parameters recommendation alert (log_duration, log_min_duration_statement, log_error_verbosity, log_statement, log_statement_stats). We will cover how to troubleshoot and resolve the issue to maintain your server's performance.
Click on each of the highlighted recommendations above. This action will direct you to a detailed page about the recommendation alert you received. On this page, you will find comprehensive information in the Recommendation details section and the reason for receiving it, which is provided in the 'Reason' column. The 'Reason' column for this recommendation will state the value of the parameter set. For example: “log_statement_stats is set to ON”.
Click on the 'Recommended Action' to proceed with the necessary steps to resolve your server's vulnerable state. The necessary action is outlined in the 'Recommended Action' section. In these scenarios, you will be directed to the server parameters page on Azure portal to update the server parameter values as provided in the “Recommended Action”.
Setting the server parameter values per the recommendation action enables your server resources to be utilized efficiently.
Scenario 7 – Long running transaction
Let us walk through a scenario where you receive a long running transaction recommendation alert. We will cover how to troubleshoot and resolve the issue to maintain your server's performance.
Click on the highlighted recommendation above. This action will direct you to a detailed page about the recommendation alert you received. On this page, you will find comprehensive information in the Recommendation details section and the reason for receiving it, which is provided in the 'Reason' column. The 'Reason' column for this recommendation states: “Here are the longest running transactions. PID(s) [311568] Duration (hours): [29]”.
Click on the 'Recommended Action' to proceed with the necessary steps to resolve your server's vulnerable state. The necessary action is outlined in the 'Recommended Action' section. In this scenario, you will need to perform the mitigation steps provided in the troubleshooting guide. Clicking on the 'Recommended Action' will take you to the troubleshooting guides. Please visit the High CPU -> Long running transactions section to find the recommended actions to resolve the issue.
Scroll to the bottom of the page to find the Recommendations to resolve the issue as shown in the screenshot below.
Follow the recommendation and terminate the session for the Pid’s identified to keep your server performant.
Scenario 8 – High CPU Utilization
Let us walk through a scenario where you receive a high CPU utilization recommendation alert. We will cover how to troubleshoot and resolve the issue to maintain your server's performance.
Click on the highlighted recommendation above. This action will direct you to a detailed page about the recommendation alert you received. On this page, you will find comprehensive information in the Recommendation details section and the reason for receiving it, which is provided in the 'Reason' column. The 'Reason' column for this recommendation states: “90% or higher CPU utilization for more than 2 consecutive hours”.
Click on the 'Recommended Action' to proceed with the necessary steps to resolve your server's vulnerable state. The necessary action is outlined in the 'Recommended Action' section. In this scenario, you will need to perform the mitigation steps provided in the troubleshooting guide. Clicking on the 'Recommended Action' will take you to the troubleshooting guides. Please visit the High CPU section to find the reasons for the CPU utilization to be high and take recommended actions to resolve the issue.
Review each of the highlighted tabs in the screenshot to identify the issue. Each tab contains recommendations to resolve the specific causes of high CPU utilization. As there might be multiple reasons, make sure to go through all the tabs and follow the recommendations provided accordingly.
Conclusion:
Imagine having a proactive ally that helps you stay ahead of potential issues with your server; That's exactly what Azure Advisor offers! In many scenarios, Azure Advisor recommendations will empower you to clearly identify and depict your server's vulnerable state. Not only does it highlight these vulnerabilities, but it also provides you with actionable recommendations to ensure your server remains highly performant. With Azure Advisor by your side, you can confidently optimize your server's performance and security, making your operations smoother and more efficient.
Updated Feb 28, 2025
Version 2.0
Gayathri_Paderla
Microsoft
Microsoft
