Managed Alerts for Adobe Commerce to Check the Health of Your Site

Creating and maintaining amazing websites is the need of modern businesses. Hence, there is a consistent search for the best tools and technologies to deliver the best for business websites. For e-commerce businesses, Magento is the best open-source e-commerce platform in PHP. Its enterprise version, or Magento Open Source, is now Adobe Commerce after the acquisition of Magento by Adobe.

Adobe Cloud Commerce Development or Magento Open Source is a fully managed cloud-based Magento version fully integrated with Adobe tools. It is the perfect all-in-one tool when it comes to designing online businesses aiming to expand their growth and offer a high-end customer experience.

Out of all the useful features of Adobe Commerce, like upsell, cross-sell, customer segmentation, content management, marketing tools, business intelligence dashboards, customer loyalty tools, etc., developers rely highly on managed alerts. Let us know more about these alerts, followed by a quick list of the top warning and critical alerts.

How to Manage Alerts for Adobe Commerce?

Websites reaching critical storage or Apdex levels (user satisfaction with services and application response time) need immediate attention. A small effort on time can save the e-commerce platform from outages or slow response time. Adobe Commerce has a solution for cloud infrastructure projects.

It allows access to the three New Relic services to monitor, manage, and troubleshoot the infrastructure and application. In addition, these New Relic services collect, view, and analyze data from Adobe Commerce on cloud infrastructure projects. 

• New Relic APM- Pro and Starter
• New Relic Infrastructure- Pro and Production environment only
• New Relic Logs- Pro, Production, and Staging environment

Hence, the alert policy is made available after the notification channel setup in New Relic services. Therefore, any developer-facing no Adobe Commerce alert policy can be due to the recently configured New Relic services. The alert policy is made available the next day after running the process.

For receiving managed alerts to understand the health of your site, use Adobe Commerce on cloud infrastructure Pro plan architecture. You will receive Apdex and error rate conditions alerts if you’re Adobe Commerce on cloud infrastructure Starter plan architecture.

Let us have a quick look at the managed alerts for Adobe Commerce to check the health of your site in table form, followed by a detailed analysis of every alert.

Types of Alerts in Adobe Commerce

After having a quick look at the managed alerts for Adobe Commerce to check the health of your site, it comes to the details of these alerts. 

It is essential to note here that the set thresholds for warning and critical alerts are based on the historical performance data across the customer base. Further, these are supported by the latest insights from Adobe Commerce’s Support and Engineering teams. Hence, all these thresholds can be changed based on ongoing and latest analysis.

The typical alert flow for these managed alerts is from lower to higher seriousness. Hence, a warning alert is raised before the critical alert, and so on. All these alerts are developed by Adobe to share a standard set using insights to the customers from the Support and Engineering. Thus, whenever one or more threshold is passed, the respective alert comes up. Let us try to find out about these managed alerts in detail.

1. CPU Warning Alert

Issue:

Based on the selection of the alert notification channel, the CPU warning alert looks like the following:

The quick do’s and don’ts for the CPU warning alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to immediately put the website into maintenance mode if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The Steps for Identifying and Troubleshooting the Cause of Cpu Warning Alert Are:

1. Start by using the New Relic APM’s Transaction page to identify transactions with performance issues like:

• Sort transactions by ascending Apdex scores. A low Apdex score indicates an issue.
• Sort transactions by threshold limits, most time-consuming, slowest average response time, highest throughput, etc.

• If the source is unknown, it is easy to use the New Relic APM’s Infrastructure page to find the resource-heavy services.

3. After identifying the source, the next step is to SSH into the environment for a detailed investigation.

4. The other steps that can work in source identification are:

• Go through the recent trends to identify any recent code deployment or configuration changes.
• Go for checking and disabling any flat catalogs.
• Try blocking the bot traffic if facing a DDoS attack.

5. Use the mitigation steps like upsizing or placing the website into maintenance mode if the problem seems temporary. Consider a permanent upsize if it returns the website to normal operations.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture.

2. Cpu Critical Alert

Issue:

Based on the selection of the alert notification channel, the CPU critical alert looks like the following:

The quick do’s and don’ts for the CPU critical alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to immediately put the website into maintenance mode if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of CPU critical alert are:

1. Start by confirming the ticket information by selecting “New Relic CRITICAL alert received” and go through the alert description.

2. Use the using the New Relic APM’s Transaction page to identify transactions with performance issues like

• Sort transactions by ascending Apdex scores. A low Apdex score indicates an issue.
• Sort transactions by threshold limits, most time-consuming, slowest average response time, highest throughput, etc.

• If the source is unknown, it is easy to use the New Relic APM’s Infrastructure page to find the resource-heavy services.

4. After identifying the source, the next step is to SSH into the environment for a detailed investigation.

5. Use the mitigation steps like upsizing or placing the website into maintenance mode if the problem seems temporary. Consider a permanent upsize if it returns the website to normal operations.

6. You can try to reproduce the problem in the Dedicated Staging by running the load test and optimizing the code to reduce pressure on the services.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture.

3. Memory Warning Alert

Issue:

Based on the selection of the alert notification channel, the memory warning alert looks like the following:

The quick do’s and don’ts for the memory warning alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of memory warning alert are:

1. Start by looking for memory-intensive processes using the New Relic APM’s Infrastructure Page. It is easy to go for the following steps if MySQL or Redis is consuming memory:

2. Confirm if MySQL or Redis is due for some update or not. The latest versions are optimized for resource consumption.

3. It is easy to check MySQL for duplicate indexes, primary keys not defined, long-running queries, etc.

4. If there are no MySQL issues, it can be PHP causing memory-intensive processes. It can be checked by running the “ps aufx” command in the CLI/Terminal.

5. Start by using the New Relic APM’s Transaction page to identify transactions with performance issues like

• Sort transactions by ascending Apdex scores. A low Apdex score indicates an issue.
• Sort transactions by threshold limits, most time-consuming, slowest average response time, highest throughput, etc.

• If the source is unknown, reviewing the past seven days of the activity is recommended for any correlations in code deployments or changes. 
• If you’re still unable to find the source, requesting an upsize or placing the website in maintenance mode is easy.

8. Consider a permanent upsize if it returns the website to normal operations.

9. You can further try to reproduce the problem in the Dedicated Staging by running the load test and optimizing the code to reduce pressure on the services.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture.

4. Memory Critical Alert

Issue:

Based on the selection of the alert notification channel, the memory critical alert looks like the following:

The quick do’s and don’ts for the memory critical alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of memory critical alert are:

1. Start by confirming the ticket information by selecting “New Relic CRITICAL alert received” and go through the alert description.

2. If the source is unknown, it is easy to use the New Relic APM’s Infrastructure page to find the resource-heavy services.

Start by looking for memory-intensive processes using the New Relic APM’s Infrastructure Page. It is easy to go for the following steps if MySQL or Redis is consuming memory:

1. Confirm if MySQL or Redis is due for some update or not. The latest versions are optimized for resource consumption.

2. It is easy to check MySQL for duplicate indexes, primary keys not defined, long-running queries, etc.

3. If there are no MySQL issues, it can be PHP causing memory-intensive processes. It can be checked by running the “ps aufx” command in the CLI/Terminal.

4. Use the using the New Relic APM’s Transaction page to identify transactions with performance issues like

• Sort transactions by ascending Apdex scores. A low Apdex score indicates an issue.
• Sort transactions by threshold limits, most time-consuming, slowest average response time, highest throughput, etc.

• If the source is unknown, reviewing the past seven days of the activity is recommended for any correlations in code deployments or changes. 
• If you’re still unable to find the source, requesting an upsize or placing the website in maintenance mode is easy.

5. Consider a permanent upsize if it returns the website to normal operations.

6. You can further try to reproduce the problem in the Dedicated Staging by running the load test and optimizing the code to reduce pressure on the services.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture.

5. Apdex Critical Alert

Issue:

Based on the selection of the alert notification channel, the Apdex critical alert looks like the following:

The quick do’s and don’ts for the Apdex critical alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of the Apdex critical alert are:

1. Start by confirming the ticket information by selecting “New Relic CRITICAL alert received” and go through the alert description.

2. Start by using the New Relic APM’s Transaction page to identify transactions with performance issues like

• Sort transactions by ascending Apdex scores. A low Apdex score indicates an issue.
• Sort transactions by threshold limits, most time-consuming, slowest average response time, highest throughput, etc.

3. Start by looking for memory-intensive processes using the New Relic APM’s Infrastructure Page. It is easy to go for the following steps if MySQL or Redis is consuming memory:

• Confirm if MySQL or Redis is due for some update or not. The latest versions are optimized for resource consumption.
• It is easy to check MySQL for duplicate indexes, primary keys not defined, long-running queries, etc.
• If there are no MySQL issues, it can be PHP causing memory-intensive processes. It can be checked by running the “ps aufx” command in the CLI/Terminal.

4. After identifying the source, the next step is to SSH into the environment for a detailed investigation.

• If the source is unknown, reviewing the past seven days of the activity is recommended for any correlations in code deployments or changes. 
• If you’re still unable to find the source, requesting an upsize or placing the website in maintenance mode is easy.

7. Consider a permanent upsize if it returns the website to normal operations.

8. You can further try to reproduce the problem in the Dedicated Staging by running the load test and optimizing the code to reduce pressure on the services.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture and Adobe Commerce on cloud infrastructure Starter plan architecture.

6. Apdex Warning Alert

Issue:

Based on the selection of the alert notification channel, the Apdex warning alert looks like the following:

The quick do’s and don’ts for the Apdex warning alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of Apdex warning alert are:

1. Start by using the New Relic APM’s Transaction page to identify transactions with performance issues like

• Sort transactions by ascending Apdex scores. A low Apdex score indicates an issue.
• Sort transactions by threshold limits, most time-consuming, slowest average response time, highest throughput, etc.

2. If the source is still not known, it is easy to use the New Relic APM’s Infrastructure page to find the resource-heavy services by looking for memory-intensive processes using the New Relic APM’s Infrastructure Page. It is easy to go for the following steps if MySQL or Redis is consuming memory:

3. Confirm if MySQL or Redis is due for some update or not. The latest versions are optimized for resource consumption.

4. It is easy to check MySQL for duplicate indexes, primary keys not defined, long-running queries, etc.

5. If there are no MySQL issues, it can be PHP causing memory-intensive processes. It can be checked by running the “ps aufx” command in the CLI/Terminal.

6. After identifying the source, the next step is to SSH into the environment for a detailed investigation.

• If the source is unknown, it is recommended to review the past seven days of the activity for any correlations in code deployments or changes. 
• If you’re still unable to find the source, it is easy to request an upsize or place the website in maintenance mode.

7. Consider a permanent upsize if it returns the website to normal operations.

8. You can further try to reproduce the problem in the Dedicated Staging by running the load test and optimizing the code to reduce pressure on the services.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture and Adobe Commerce on cloud infrastructure Starter plan architecture.

7.  Disk Warning Alert

Issue:

Based on the selection of the alert notification channel, the disk warning alert looks like the following:

The quick do’s and don’ts for the disk warning alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of disk warning alert are:

1. Start by reviewing the disks for the highest usage. If there is a slow increase in the disk usage, you can try the following options:

• Optimize the disk space by adjusting space allocation.
• Clear up the space for MySQL.
• If the New Relic shows the increased disk usage, which indicates problems causing the file to increase rapidly in the directory.  It is easy to try the following options:

2. Check the overall disk space to identify the problems by running “df- h” in the CLI/ Terminal.

3. Check the attached file system after identifying the increased disk usage. It is easy to check the file directory by using the command “pub/media/.”

4. It is easy to remove the directories that are rapidly increasing the disk usage.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture.

8.  Disk Critical Alert

Issue:

Based on the selection of the alert notification channel, the disk critical alert looks like the following:

The quick do’s and don’ts for the disk critical alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of disk critical alert are:

1. Start by confirming the ticket information by selecting “New Relic CRITICAL alert received” and go through the alert description.

2. Start by reviewing the disks for the highest usage. If there is a slow increase in the disk usage, you can try the following options:

• Optimize the disk space by adjusting space allocation.
• Clear up the space for MySQL.

3. If the New Relic shows the increased disk usage, which indicates problems causing the file to increase rapidly in the directory.  It is easy to try the following options:

4. Check the overall disk space to identify the problems by running “df- h” in the CLI/ Terminal.

5. Check the attached file system after identifying the increased disk usage. It is easy to check the file directory by using the command “pub/media/.”

6. It is easy to remove the directories that are rapidly increasing the disk usage.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture.

9. MariaDB Alerts

Issue:

Based on the selection of the alert notification channel, MariaDB alerts look like the following:

The quick do’s and don’ts for the MariaDB alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.
• It is necessary to end any scripts like imports causing the site performance issues.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.

Solution:

It is essential to note here that if you receive a DML Queries Critical alert in step one, it is followed by the DML Queries Warning alert in the second step. The steps for identifying and troubleshooting the cause of the MariaDB alert are:

1. Start by checking if the Adobe Commerce support ticket exists with the information “New Relic MariaDB alert received.” It must have the alert description and the “New Relic Incident Link,”

2. Try identifying the issue’s source through DML queries.

3. Review the database operations by New Relic steps in the following way- “APM UI Pages> Monitoring> Databases page.” 

4. Sort by “CALL COUNT,” then “OPERATION,” and review “INSERT,” “DELETE,” and “UPDATE” operations.

5. Search high AVG.

6. Click through the database operation callers by identifying transactions using the query by time.

7. Look for code optimization or operational optimization in the following ways:

• For code optimization: Search the optimized queries with throttling code, minimized index usage, bulk updates/ inserts, etc.
• For operational optimization: Offload the resource-intensive data modifications to lower traffic times.
• Further, ensure that the system is on the latest version of ECE tools.

Affected Products and Versions:

Adobe Commerce on cloud infrastructure Pro plan architecture.

10. Redis Memory Critical Alert

Issue:

Based on the selection of the alert notification channel, the Redis memory critical alert looks like the following:

The quick do’s and don’ts for the Redis memory critical alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of the Redis memory critical alert are as follows:

1. Start by confirming the ticket information by selecting “New Relic CRITICAL alert received” and go through the alert description.

2. If there are no support tickets, it is easy to go through the Redis Used Memory increasing or decreasing by going to “one-newrelic.com,” then “Infrastructure,” and “Third-party services” page.

3. Submit the support ticket to have the cluster upsized, or you can increase the “maxmemory” limit to the next level.

4. If the source is still not known, it is recommended to review the past seven days of the activity for any correlations in code deployments or changes. 

5. It is the time to check for misbehaving third-party extensions in the following ways:

• Start by correlating the recently installed third-party extensions and the time of issue occurrence.
• Review the extensions affecting the Adobe Commerce cache and causing it to grow quickly.

6. If all the steps mentioned above are not helping, consider enabling the L2 cache to reduce network traffic between Redis and app. It is easy to enable the L2 cache for cloud infrastructure in the following ways:

• Upgrade the ECE Tools.
• Configure the L2 Cache by using “Use REDIS_BACKEND variable” and updating “.magento.env.yaml” file.

Affected Products and Versions:

All versions of the Adobe Commerce on cloud infrastructure Pro plan architecture.

11. Redis Memory Warning Alert

Issue:

Based on the selection of the alert notification channel, the Redis memory warning alert looks like the following:

The quick do’s and don’ts for the Apdex critical alert to prevent the website from becoming non-responsive are:

Do’s:

• Abort all the scheduled deployments until the alert is cleared. 
• It is necessary to put the website into maintenance mode immediately if it becomes unresponsive. You must add the IP to the exempt IP address list to ensure quick access to the website for troubleshooting purposes.

Don’ts:

• Clear the cache memory.
• Do any administrative tasks like data imports/ exports, Commerce Admin, etc.
• Run additional crons or indexes, causing additional stress on the CPU or disk.
• Launch any new marketing campaign attracting additional page views to the website.

Solution:

The steps for identifying and troubleshooting the cause of Redis memory warning alert are:

1. Check if the Redis Used Memory increasing or decreasing by going to “one-newrelic.com,” then “Infrastructure,” and “Third-party services” page.

2. Submit the support ticket to have the cluster upsized, or you can increase the “maxmemory” limit to the next level.

3. If the source is still not known, it is recommended to review the past seven days of the activity for any correlations in code deployments or changes. 

4. It is the time to check for misbehaving third-party extensions in the following ways:

• Start by correlating the recently installed third-party extensions and the time of issue occurrence.
• Review the extensions affecting the Adobe Commerce cache and causing it to grow quickly.

5. If all the steps mentioned above are not helping, consider enabling the L2 cache to reduce network traffic between Redis and app. It is easy to enable the L2 cache for cloud infrastructure in the following ways:

• Upgrade the ECE Tools.
• Configure the L2 Cache by using “Use REDIS_BACKEND variable” and updating “.magento.env.yaml” file.

Affected Products and Versions:

All versions of the Adobe Commerce on cloud infrastructure Pro plan architecture.

Conclusion

Adobe Commerce offers different managed alerts for users to take timely action regarding optimized website performance. Hence, users of Adobe Commerce on cloud infrastructure Pro plan architecture can seek the guidance of the managed alerts for Adobe Commerce to check the health of their site. After ensuring that New Relic services are enabled, it is easy to respond according to the type of managed alerts. As an Adobe Commerce development company with certified experts on board, we help you build and go live quickly with all the specifics you want in place.

All these managed alerts are received in the flow of seriousness, i.e., the first warning alert is followed by the critical alert. Developers can go through the managed alerts mentioned above for the CPU, memory, disk, Apdex, Maria DB, and Redis memory. Understanding the affected products and versions, issues, and possible solutions is easy.