Hello, this is Sudo (
id:do-su-0805) from the Mackerel CRE team.
We would like to share the details of our first update of 2026. We appreciate your continued support this year.
- You can now search by “Status” in the trace screen filter conditions
- Automatic retirement for ALB and NLB is now available in AWS Integration
- Additional metrics are now available when integrating ALB via AWS Integration
- Fixed an issue where hosts could be retired unintentionally when filtering ECS Cluster and Lambda by tags
- Display period is now shown for custom metric graphs on the host detail screen
- [Notice] Network interface information retrieval behavior will change in the Windows version of mackerel-agent
- New products have been added to mackerel-labs
You can now search by “Status” in the trace screen filter conditions
When searching for traces in APM, Status has been added to the filter conditions. You can search based on the status within spans. In the List Traces API, you can filter results using the 'statusCode' key.
Automatic retirement for ALB and NLB is now available in AWS Integration
When ALB or NLB is integrated via AWS Integration, automatic retirement is now available. With this feature, when a target resource is deleted in your AWS account, the corresponding host in Mackerel is automatically retired. You can enable automatic retirement from the AWS Integration settings screen.
For details about the automatic retirement feature, please refer to the following help page.
AWS Integration - Mackerel Docs
Additional metrics are now available when integrating ALB via AWS Integration
When ALB is integrated via AWS Integration, the following metrics are now additionally available.
- ConsumedLCUs
- alb.consumed_lcus.all
- PeakLCUs
- alb.peak_lcus.maximum
- alb.peak_lcus.sum
If Add new metrics automatically is enabled in your AWS Integration settings, collection will start automatically. If Add new metrics automatically is disabled, metrics will be collected after selecting them under Specify the metrics to be retrieved.
For details on the newly available metrics, please refer to AWS Integration - ALB - Mackerel Docs and the AWS documentation CloudWatch metrics for your Application Load Balancer - ELB
Fixed an issue where hosts could be retired unintentionally when filtering ECS Cluster and Lambda by tags
When automatic retirement is enabled in AWS Integration, the corresponding Mackerel host is retired when a target resource is deleted. We have fixed an issue where, in some cases, the corresponding host could be retired unintentionally when integrating ECS Cluster or Lambda while using tag-based filtering.
Display period is now shown for custom metric graphs on the host detail screen
On the host detail screen, system metric graphs display the metric time range next to the graph name, but custom metric graphs did not, making it difficult to confirm the display period.
This has been improved so that custom metric graphs now also display the time range next to the graph name, allowing you to easily confirm the displayed period.
[Notice] Network interface information retrieval behavior will change in the Windows version of mackerel-agent
We have identified an issue on certain versions of Windows Server where metric collection fails if the network interface device driver name contains Japanese characters. This issue was caused by a flaw in the previous implementation and will be fixed in an upcoming release of mackerel-agent. As a result of this change, when updating from an existing Windows version of mackerel-agent (version 0.85.3 or earlier) to the new version, the following effects may occur:
- The network names displayed in Host Details may change
- Network interfaces that were previously excluded may become targets for metric collection
- Network interface names and metric names may change
If you are using these metrics in monitoring rules (including usage in expression graphs), please be aware that follow-up changes may be required.
New products have been added to mackerel-labs
The following products have been added to mackerel-labs, a collection of experimental products provided to make monitoring with Mackerel more convenient.
- mackerel-plugin-supervisord
- A metric plugin that can be used to monitor Supervisor
- mackerel-plugin-snmp-table
- A metric plugin that retrieves tabular SNMP values and converts each into metrics
- sabatrapd
- Middleware that receives SNMP Traps from network devices and posts them as host check monitoring items
- sabatrafficd
- A program that queries network switches and other devices via SNMP, collects interface traffic statistics, and sends the information to Mackerel
Please follow the instructions in each README for installation and configuration. Note that software published in Mackerel Labs is experimental and not covered by support. We are generally unable to respond to inquiries regarding these products. Thank you for your understanding.
