Mackerel blog #mackerelio

The Official Blog of Mackerel

The billing system will be updated to support containers and cloud components starting May 2019

Thank you for choosing Mackerel.

Since the recent release of the public beta version of Mackerel’s container agent, the number of cases used in container environments has increased.

mackerel.io

In Mackerel’s existing billing system, all components are converted into a unit called a ‘host’ based on usage, and calculated at a rate of 1,800 yen per host. The host unit has a strong connotation as equivalent to one server, and this concept doesn’t really match actual applications with recent container environments and cloud managed services.

So, starting from May 2019, the billing system will be newly revised, dividing the existing ‘host’ billing unit into two separate billing units, standard hosts and a micro hosts.

With the transition to the new billing system, there are no operations or procedures that need to be performed by the user. The new rates will be applied starting in May 2019, thus affecting the bill of June 2019.

Below, we’ll go over some of the details regarding standard host and micro host units.

Standard hosts

Standard hosts will function similar to the current ‘host’ unit. 1 standard host will be charged 1,800 yen (tax included).

The following are billable targets for a standard host.

  • 1 physical server or virtual server installed with mackerel-agent will count as 1 standard host
  • 1 VM system integration (AWS EC2, AzureVM) instance will count as 1 standard host
  • If the number of host metric exceeds 200, every 200 host metrics will count as 1 standard host
  • Every 200 service metrics will count as 1 standard host
  • Every 20 external monitoring items will count as 1 standard host

Micro hosts

Micro hosts are a new billing unit. 1 micro host will be charged 648 yen (tax included).

The following are billable targets for a micro host.

  • 1 pod or task with mackerel-container-agent registered as a sidecar will count as 1 micro host
  • Apart from VM system integration calculated as standard hosts, every AWS/Azure Integration component will count as 1 micro host
  • If the number of host metrics for a micro host exceeds 30, every 30 host metrics will count as 1 micro host

Micro host targets that are currently being charged as regular ‘hosts’ will be automatically change to being calculated as micro hosts starting in May 2019.

If you have any questions or concerns, contact the support team at support@mackerel.io.

【Advanced release notice】Specification changes to the organization’s notification disable setting

Thank you for choosing Mackerel.

This is an advanced notice to inform you that changes will be made to the specifications of the notification disable function with the release on Thursday, April 18, 2019.

The contents of these changes follow below.

Before the changes

According to the current specifications, the following notifications will not be sent while organization’s notifications are disabled.

  • Notifications related to alerts or alert groups

f:id:mackerelio:20190410111726p:plain These specifications refer to the ‘Turn Off Notifications’ button in the above image.

After the changes

After changes are made to the specifications, the following notifications will not be sent while notifications are disabled.

  • Notifications related to alerts or alert groups
  • Notifications related to host status · new host registration · retired hosts
  • Notifications related to creating, updating, deleting monitoring settings

Graph sharing and test notifications will continue to be unaffected by disabling notifications.

Intention behind the specification changes

To clarify, targets of the notification disable function, previously only alert and alert group related notifications, will be updated to all notification events.

Although this function has a wide range of influence, using it will make it difficult to know who changed what and for what reason, and because of this, we do not recommend disabling notifications. In addition, the functions location at the top of the monitoring settings made it difficult to understand its influence when configured.

With this specification change, all notification events will be targeted and the disable button will be moved to the notification channel’s list page. Please keep in mind that configuring this function has a wide range of influence.

Thank you for your cooperation.

User information added to the payload of Webhook notifications and more

Mackerel Team CRE Miura (id:missasan) here.

The long Golden Week holiday is fast approaching and I am excited. The Mackerel business team will be participating at two events surrounding the extended holiday, the Gartner Conference 2019 and the Cloud Computing EXPO【Spring】. If you’re planning on attending one of these events, be sure to stop by the Mackerel booth. We have new Mackerel goods to give away!

Now on to this week’s update information.

User information added to the payload of Webhook notifications

User information has been added to the payload of Webhook notifications for when monitor changes are made. This change makes it possible to respond to requests such as wanting automatic version management for mkr monitor content triggered by a change being made.

Sorting by id now included for mkr monitors command

In order to prevent inconsistencies in sorting order when there are monitors of the same type and name, the sorting conditions have been modified to include id. Note that after this change, there may be a diff output in the first mkr monitors diff result executed even if there are no changes to the monitoring rule.

Mackerel at the Cloud Computing EXPO【Spring】from May 8-10!

The Mackerel team will be running a booth at the Cloud Computing EXPO Spring to be held at Tokyo Big Sight from May 8th until the May 10th.

If you're going to the event, be sure to stop by the Mackerel booth.

spring.japan-it.jp

Event Details

ECS now supported with AWS Integration and more

Mackerel Team CRE Miura (id:missasan) here.

Earlier this year in February, the public beta version of Mackerel’s container agent was released. And with the growing interest in container monitoring, AWS Integration has now been updated to support ECS. With this feature, you can integrate and check the metrics of ECS clusters and services in Mackerel. Definitely give it a try with the container agent.

A large number of OSS updates were made this week as well. Thank you to all the contributors!

Now on to this week’s update information.

ECS now supported with AWS Integration

For details regarding obtainable metrics and billing information, refer to the following help page.

mackerel.io

[go-check-plugins] check-ping added

With the release of go-check-plugins v0.29.0, check-ping was added. This can be executed by specifying something along the lines of check-ping -H 127.0.0.1 -n 5 -w 100. Using this plugin, you can now monitor the connectivity of network devices that can’t be installed with Mackerel agent.

For more details on configuration methods, check out the README linked below.

github.com

[go-check-plugins] Option added to exclude service names from check-ntservice

With the release of go-check-plugins v0.29.0, we also saw the addition of the --exclude-service option which allows you to specify a service name to be excluded in check-ntservice.

github.com

[go-check-plugins] Option added to check-ntpoffset to check whether the NTP upper stratum is correctly synchronized

The release go-check-plugins v0.29.0 also added the --check-stratum option which allows you to check whether or not the NTP upper stratum is correctly synchronized in check-ntpoffset .

github.com

[go-check-plugins] Option added to specify the http proxy with check-http

With the release of go-check-plugins v0.29.0, the --proxy option was added to check-http to specify the http proxy. Before now, this could be specified in environment variables, but is now even easier to use as it can be expressed in the command line option.

github.com

[mkr] mkr monitors now supports threshold configuration for service metric interruption monitoring

With the release of mkr v0.36.0, the mkr monitors command now supports threshold configuration for interruption monitoring of service metrics.

github.com

Mackerel at the Gartner Conference 2019 from April 23rd until April 25th!

Mackerel will be running a booth at the Gartner IT Infrastructure, Operations & Cloud Strategies Conference 2019 held in the the main building of Happo-en from April 23rd (Tue) until April 25th (Thu). If you’re planning to attend the event, be sure to stop by the Mackerel booth.

www.gartner.com

Event Details

Experimental Feature: Expression Monitoring was stopped from 4/2 to 4/3

Thank you for choosing Mackerel.

Expression Monitoring, which is offered as an Experimental Feature, was stopped during the following time period.

  • April 2 (Tuesday) 11:54 a.m. - April 3 (Wed) 5:49 p.m. (JST)

This stoppage was caused by a misconfiguration of the server that performs batch processing in relation to the Expression Monitoring feature. We apologize for the inconvenience and thank you for your understanding.

Monitoring for service metric interruption is now possible and more

Mackerel Team CRE Miura (id:missasan) here.

Due to popular demand, we’ve released a function for service metric interruption monitoring.

Service metrics are versatile because they can be posted using the API without having to go through the Mackerel agent. With this release, it is now possible to detect when service metric posting gets interrupted and prevent the situation from going unnoticed.

Now on to this week’s update information.

Monitoring for service metric interruption is now possible

With Mackerel, you can post and visualize metrics from external services and applications that aren’t directly linked to a specific server using the service metric API. You can do this by hitting the API directly or with fluentd.

Interrupted monitoring for service metrics is a mechanism to detect whether the process of posting these metrics gets interrupted for some reason.

Sometimes stoppages occur with fluentd, and we’ve received feedback from users saying that it took them a few days to even notice that service metrics weren’t being sent. This kind of situation can now be prevented by configuring interruption monitoring.

Configuration Method

Configuring interruption monitoring can be done from the service metric monitoring tab in Monitor Settings or from the editing screen of existing service metric monitors.

f:id:mackerelio:20190322182242p:plain

Graph legend display for Graphboard expression graphs

As shown in the image below, a graph legend can now be displayed on the bottom of expression graphs that are on the Graphboard which can be created from the service details screen.

f:id:mackerelio:20190325142134p:plain

Release of mkr v0.35.1

Hello! Mackerel Team CRE id:a-know here.

The weather has warmed up considerably and I’m beginning to see some cherry blossoms popping up here and there on my commute to work. Anyways, on to this week's release contents.

Release of mkr v0.35.1

With the release of version 0.35.1 for Mackerel’s command line tool mkr, the following bug contained in version 0.35.0 has been fixed.

  • The specification content for the hosts subcommand --status option not being applied.

Just to be safe, we recommend checking what is installed on your PC and the version of mkr being used with regularly run scripts.

github.com