Mackerel

Mackerel is a cloud monitoring service for teams responsible for servers, services, and container workloads. It is best suited to organizations that want infrastructure and business-facing metrics in one hosted system. Its coverage and extensibility are compelling, but usage-based charges require disciplined planning as monitoring expands.

Overview

Installing Mackerel’s agent begins host monitoring, after which hosts can be arranged into services and roles. This structure maps machine signals to the applications and teams that own them, rather than leaving operators with an undifferentiated server list. User-defined service metrics can track response time, sales, page views, and other measures alongside host data.

Mackerel can combine metrics, monitor predicted values, and apply machine-learning anomaly detection. Amazon Web Services, Microsoft Azure, and Google Cloud integrations cover the major public-cloud environments. Its container agent treats an Amazon ECS task or Kubernetes pod as a host, extending the same monitoring model to orchestrated workloads. Open-source plugins, APIs, and the mkr command-line tool provide several routes for adapting the service to local systems.

Key features

Host, service, and container monitoring

The combination of VM monitoring, service monitoring, and container support is Mackerel’s central advantage. Teams can relate infrastructure health to application and operational indicators in one cloud service. Treating ECS tasks and Kubernetes pods as hosts is straightforward conceptually, although host-based billing makes large or highly dynamic container estates worth modeling carefully before rollout.

Metrics and detection

Custom metrics accommodate signals that generic infrastructure dashboards miss. Combined metrics and predicted-value checks help identify relationships and expected trends, while machine-learning anomaly detection can surface unusual behavior without relying solely on fixed thresholds. These capabilities are more useful to teams with meaningful historical data and clear ownership of the signals than to a small deployment needing only basic uptime checks.

Extensibility and assistance

Plugins, APIs, and mkr make Mackerel suitable for environments with bespoke tooling or deployment processes. Customer Reliability Engineers can provide technical guidance through implementation and use. Standard-plan technical support is email-only, with free contact during business days from 10:00 to 19:00, so teams needing around-the-clock support should account for that limitation.

Pricing

Mackerel uses a freemium model. The Free plan costs nothing and allows five standard and micro hosts in total, 500 metrics, one day of data retention, 10 monitoring rules, one custom dashboard, and up to 5M trace spans. That is enough for a small proof of concept or a tightly bounded environment, but the short retention and low host, rule, and dashboard allowances make it unsuitable as a broad historical monitoring system.

Standard is custom-priced at usage rates: 2,180 yen (tax included) per standard host month, 660 yen (tax included) per micro host month, 11 yen (tax included) per metric, and 330 yen (tax included) per 1M spans. It retains data for 460 days, includes 100 monitoring rules and 100 custom dashboards, and supports unlimited users. A minimum usage fee applies. The separate Standard micro-host option supports 30 metrics per micro host, making it a possible fit for lightly instrumented instances, but it does not remove the need to track host counts, metrics, and trace volume.

A 14-day trial provides Standard-plan functions and includes 100M trace spans. Use it to estimate the recurring effect of your actual host, metric, and span inventory before committing. The paid pricing note starts at 2,180 JPY per month, while the detailed rates above determine how a larger deployment scales.

Platforms

Mackerel is delivered as a cloud service with API, Linux, and Windows coverage. It supports agent-based monitoring and agentless monitoring, plus VM and service monitoring. Linux and Windows support covers common server estates, while the API and cloud integrations help connect managed infrastructure without limiting the service to one provider.

Who it's for

Mackerel fits infrastructure and platform teams that need a shared view of hosts, applications, cloud resources, and containers, especially when custom operational or business metrics matter. Its role-and-service organization and unlimited Standard users suit collaborative monitoring. The Free plan is appropriate for evaluation or a very small estate. Organizations with rapidly changing host populations, heavy trace volumes, or strict cost ceilings should model usage closely; a simpler fixed-scope monitor may be easier to budget.

Pros and cons

Pros

• Broad infrastructure coverage: Hosts, services, AWS, Azure, Google Cloud, ECS, and Kubernetes fit one monitoring model.
• Useful context beyond servers: Custom service metrics can place response time and business indicators beside infrastructure data.
• Strong extension options: Open-source plugins, APIs, and the mkr CLI support customized workflows.
• Team-oriented scale: Standard includes unlimited users, 100 rules, and 100 dashboards.

Cons

• Usage-based cost exposure: Standard billing charges separately for hosts, metrics, and trace spans, with a minimum usage fee.
• Free-plan constraints: Five hosts, 500 metrics, one-day retention, 10 rules, and one dashboard limit serious long-term use.
• Support-hours boundary: Standard technical support is email-only and limited to business-day contact hours.

Alternatives

Netdata is a better fit when a free Community plan capped at five active connected nodes and one active custom dashboard per room matches your scale, with broader device and self-hosted platform coverage.

Choose Grafana Cloud Application Observability if an always-free allowance of 2,232 host hours or a Pro model based on host-hour usage and a monthly platform fee better suits your budget.

Checkmk suits teams wanting a forever-free Community edition with more than 2,000 integrations and roughly 100 hosts, supported by the community.

Pick VictoriaMetrics when open-source single-node or cluster deployments are the priority.

Nagios XI offers a free tier for seven nodes or 100 services with standard features and self-support, while paid Standard licensing is available for larger needs.

Prometheus is the choice for a free, Apache 2.0-licensed open-source monitoring and alerting toolkit without a hosted-service billing model.

Akamai Cloud Longview is worth considering for up to 10 clients with five-minute resolution and 12 hours of history on its free tier.

Pixie provides a free hosted version of its open-source project for teams seeking that specific observability approach.

For broader category comparisons, see Cloud Monitoring and Infrastructure Monitoring.

Verdict

Mackerel is a strong choice for teams that want organized host, service, cloud, and container monitoring with custom metrics and multiple extension paths. Choose it when one collaborative cloud platform and advanced detection justify metered billing. Look elsewhere when you need highly predictable costs, very long free retention, or support outside Standard’s business-day email window.