Pakyas and Healthchecks.io both monitor cron jobs and scheduled tasks using execution signals (pings) emitted by the job itself, rather than polling a server from outside. They share the same core "dead man's switch" idea but differ in their state model, tooling, and licensing — this page lays out the differences so you can pick the right fit.
| Dimension | Pakyas | Healthchecks.io |
|---|---|---|
| Signal model | Execution-signal (heartbeat) based: the job pings on start, success, or failure. Pings hit a global edge endpoint (ping.pakyas.com/{public_id}, with /start and /fail modifiers and exit-code reporting) served by a Cloudflare Worker, then flow through a Redis stream to PostgreSQL. Optional run pairing and duration capture. | Execution-signal (heartbeat) based: jobs ping a unique URL, with separate start, success, and fail signals. Healthchecks.io originated and popularized this dead man's switch model for cron monitoring. |
| State vocabulary | Job-accurate, non-server states: On Schedule, Late, Missing, Running, Overrunning, Error, Paused, and Waiting for first ping. Deliberately avoids up/down server language — Late, Missing, and Error are distinct conditions rather than a single "down". | Five states: New (no pings yet), Up (last expected ping arrived on time), Late (a ping is overdue but still inside the check's grace time), Down (grace time elapsed without a ping — the alerting state), and Paused. The core vocabulary is server-style up/down; the Management API reports the within-grace state as 'grace'. |
| CLI tooling | First-class official CLI ('pakyas-cli', a separate Rust binary) that wraps a command, sends start/success/fail signals automatically, and reports exit codes and duration. | No official first-party CLI. Command wrapping is handled by third-party tools listed in their docs — most notably runitor (by bdd), which runs a command, sends /start then /success or /fail by exit code, and can schedule with -every. Mature, but community-maintained rather than vendor-shipped. |
| Alert deduplication / flap handling | Configurable thresholds before alerting: alert-after-N missed pings and alert-after-N failed pings, with consecutive-miss and consecutive-fail counters so transient flaps don't immediately page. Recipients are deduped by (channel type, destination). | Alerts fire once a check's grace time elapses (the Up→Down transition); the grace time is effectively the only flap buffer — there is no documented 'alert only after N consecutive failures' threshold. Repeat/nag notifications are offered instead: hourly or daily email reminders while a check is down, plus Pushover 'Emergency' re-alerts and escalation via incident-management integrations. |
| Pricing | Hosted SaaS with four tiers: Free ($0), Developer ($9/mo), Pro ($29/mo), and Business ($99/mo). Free Pro subscriptions are offered to open-source projects. | Four tiers: Hobbyist (free — monitor 20 jobs, 100 log entries per job), Supporter ($5/mo — optional financial support at the same limits), Business ($20/mo — 100 jobs, 1000 log entries per job, plus SMS/WhatsApp and phone-call credits), and Business Plus ($80/mo — 1000 jobs). Team members are unlimited on all plans; open-source projects and nonprofits can get the Business plan free on request. |
| Open-source / self-hosted | Hosted SaaS only; Pakyas does not offer an open-source or self-hostable edition. On this dimension Healthchecks.io has the clear advantage. | Fully open-source and self-hostable: the service is written in Python/Django, source at github.com/healthchecks/healthchecks under the BSD 3-Clause license, with official Docker images for running it yourself. |
| Integrations | Built-in delivery channels: Email, Slack, Discord, Telegram, Webhooks, and SMS. | 30+ notification integrations, including Email, Webhooks, Slack, Discord, Telegram, Microsoft Teams, Signal, Matrix, Mattermost, Google Chat, Zulip, PagerDuty, Opsgenie, Splunk On-Call, Pushover, ntfy, Gotify, SMS, WhatsApp, phone call, Prometheus, and Trello. |
The biggest, most credible difference is licensing and operations: Healthchecks.io offers an open-source, self-hostable edition, so if running the monitor yourself (or auditing/extending the source) matters, that is a real advantage Pakyas's hosted-only model does not match. Against that, Pakyas leans into a more opinionated, job-accurate state model — distinguishing Late, Missing, Overrunning, and Error rather than a single "down" — plus a first-class CLI and configurable flap dampening to reduce alert noise. Both tools use the same fundamental heartbeat approach, so for many teams the decision comes down to whether you value self-hosting and a mature open-source community (Healthchecks.io) or a managed service with a richer state vocabulary, integrated CLI, and edge ingestion (Pakyas). Healthchecks.io also has far broader notification-integration coverage, so if you need niche channels out of the box that weighs in its favor too.
New to the terminology? See the cron monitoring glossary for plain-language definitions, or explore everything Pakyas tracks on the features page.
Ready to try Pakyas for yourself?
Start monitoring your cron jobs with execution-signal precision — free for up to 10 checks, no credit card required.
Start monitoring free