Comparison

kogut combines server, container, and uptime monitoring. Analytics, JavaScript errors, and deploy reports add context for web apps. It runs as one Rust binary with one SQLite file.

The projects below have different priorities. Their features change, so check their current documentation before migrating.

Uptime Kuma

Uptime Kuma focuses on uptime checks and has a wider range of monitor types and notification providers. It suits setups where endpoint checks are the main requirement.

kogut adds host metrics and deploy context to uptime checks. Checks can run from an agent inside a private network. Configuration lives in the panel or in kogut.toml. See Coming from Uptime Kuma for the differences.

Beszel

Beszel monitors server and container metrics with a hub and agents. Its agents also run network monitors over ICMP, TCP, HTTP, and DNS. These record response time and packet loss, and HTTPS monitors watch certificate expiry. Beszel alerts when a monitor’s packet loss over the last hour passes a threshold. It has no status page, and its monitors run only from the monitored hosts, not from outside them. If those metrics and monitors are all you need, a focused tool means fewer features to configure.

kogut runs each check from the server or, with from, from an agent inside your network. HTTP checks assert on status codes, body text, JSON values, and response time. Failing checks, host metrics, and deploys meet in one incident with likely causes, which sends a message when a service goes down and when it recovers. A failing check, a memory spike, and a container restart become one incident with a ranked likely cause, not three messages; see the example. A status page shows service status to the public. See Coming from Beszel.

Checkmate

Checkmate combines uptime and infrastructure monitoring. It has more monitor types, such as PageSpeed, gRPC, WebSocket, and game servers, and 13 notification channels, including SMS. Its interface comes in 19 languages, and its status pages have themes and custom CSS. Its Capture agent also runs on Windows and macOS.

Checkmate runs as a Node.js server with MongoDB; kogut runs as one binary with one SQLite file. Checkmate’s server pulls metrics from Capture over HTTP on a port the agent listens on, protected by one shared secret. kogut’s agent connects out over HTTPS, buffers up to one hour or 20 MiB while the server is unreachable, and uses its own host token, which the server stores only as a hash. Checkmate opens an incident per monitor. kogut groups related failures into one incident with a likely cause.

maintenant

maintenant monitors containers, endpoints, heartbeats, and certificates on a Docker host. Its free Community edition watches only the host it runs on, with no remote agents, up to 10 endpoints and 5 heartbeats, and 7 days of history. Email, Telegram, and Slack channels and maintenance windows are paid features. It sends telemetry unless you opt out. It has no built-in sign-in by design and expects an authenticating reverse proxy in front of it.

kogut has no paid edition and no agent limit, and keeps host history as set in [retention]. It signs people in with passwords, OpenID Connect, or forward auth, and sends no telemetry.

Gatus

Gatus reads its endpoints and conditions from a YAML file and shows the results on a read-only dashboard. It collects no host metrics.

kogut takes checks from the panel or from kogut.toml, and adds agents for host metrics and checks inside private networks. See Coming from Gatus.

Umami

Umami focuses on web analytics and has more analytics reports. It is a better fit when analytics is the main task rather than context for monitoring.

kogut’s script collects page views, visitors, Web Vitals, and JavaScript errors. Its analytics have no goals or funnels. See Coming from Umami.

Tianji

Tianji combines analytics, uptime monitoring, server status, and status pages. Its scope also includes application telemetry and surveys.

kogut centres on monitoring services and relating their failures. For its resource use, see What is kogut.

Kener

Kener focuses on status pages, with monitors behind them, badges, and widgets to embed status on other sites.

kogut’s status page is one part of server, container, and uptime monitoring.

openstatus

openstatus focuses on status pages and checks from many regions, and is mainly run as a hosted service.

kogut runs each check from one place: the server or one of your agents. It adds host metrics and deploys to the same incident.

OneUptime

OneUptime covers a wider observability scope, including on-call schedules, logs, traces, and APM.

kogut leaves those out to stay small; see Limits of scope.

Betterlytics

Betterlytics focuses on web analytics and adds session replay, funnels, and uptime checks.

kogut’s analytics are context for monitoring, without session replay or funnels.

What each tool needs to run

Besides its own process, each tool needs the storage and services below, according to its documentation or issue tracker. Resource use compared measures memory, CPU, and disk for several of them.

Tool	Storage and other services
kogut	None. One SQLite file; see resource use
Uptime Kuma 2	SQLite, or MariaDB inside the container or external (README)
Beszel	None. The hub embeds PocketBase with SQLite (documentation)
Checkmate	MongoDB, which issues report failing on CPUs without AVX and on older ARM boards (#1386, #3903)
maintenant	None. SQLite, or PostgreSQL (README)
Gatus	None. Results stay in memory, or in SQLite or PostgreSQL (storage)
Umami	PostgreSQL (installation)
Tianji	PostgreSQL (README)
Kener 4	Redis, plus SQLite, PostgreSQL, or MySQL (#604)
openstatus	Self-hosted with eight containers, including libSQL and a local Tinybird (DOCKER.md); an issue lists the external services it relies on (#2905)
OneUptime	PostgreSQL, ClickHouse, and Redis; users report 14 to 16 GB of memory (#2252)
Betterlytics	ClickHouse and PostgreSQL (repository)