Application heartbeat infrastructure

Know your applications are alive. Reach them when you need to.

HeartbeatHook combines outbound heartbeat monitoring with a secure signal relay. Applications report that they are alive, and receiving applications retrieve pending signals when they are ready.

HMAC-authenticated Replay-aware requests Retry-safe signal creation
signal / contract explicit
Signal pathPending work is explicit
Tier 1 API
RequestHMACverified
MailboxPendingpoll when ready
TransitionACKexplicit
✓Heartbeat receivedapplication reports from its own runtimeALIVE
→Signal availablereceiver polls its authenticated mailboxPENDING
•Result confirmedreceiver ACKs after its business actionACKED
✓Heartbeat receivedOutbound evidence of life
Signal retrieved by pollingAuthenticated read · explicit ACK
A clear contract for application-to-application work
HMACREPLAY-AWAREIDEMPOTENTACKNOWLEDGED

A signal you can reason about

From heartbeat to ACK, every step has a clear job.

One application reports outbound evidence of life. Another can retrieve a pending signal through an authenticated polling flow and confirm the result after its own business action.

01◉

Monitor

Applications and scheduled jobs send authenticated heartbeats from the runtime that needs observing.

application → HeartbeatHook
02→

Reach

A source application creates a pending signal for another application in the same tenant.

POST /heartbeat/create
03✓

Confirm

The receiving application polls, performs its action and ACKs the signal. Repeated ACKs remain safe.

GET /heartbeat/signals · ACK

Made for serious systems

Less guessing. More signal.

HeartbeatHook keeps authentication, replay protection, idempotency, pending state and acknowledgement explicit. It is a focused application contract, not a general-purpose event platform.

Read the API contract
◉

Authenticated by default

Requests use HMAC-SHA256, timestamps, fresh nonces and request IDs.

↻

Retries without duplicate creation

Stable business idempotency keys make an uncertain create response safe to retry.

✓

Pending work stays visible

Signals remain pending until the authenticated target reads and ACKs them.

⚙

Bounded by design

Tier 1 documents request size, polling and rate-limit boundaries instead of promising unmeasured scale.

Platform products

Choose what your application needs.

Start with the operational question: do you need evidence of life, missing-check-in reasoning, progress freshness or a small signal path?

Current Tier 1 foundation

Heartbeat Monitor

Are you checking in?

Outbound evidence of life from an application, job or service.

Explore Heartbeat Monitor
Product concept

Dead Man Monitor

Did you stop checking in?

Expected heartbeat silence as a distinct monitoring condition.

Explore Dead Man Monitor
Product concept

Dead Hand Monitor

Are you checking in but stuck?

Progress-freshness monitoring for alive processes that stop moving work forward.

Explore Dead Hand Monitor
Current Tier 1 foundation

Signal Relay

Can another application retrieve it?

A bounded authenticated mailbox with explicit reads and ACKs.

Explore Signal Relay
Product layer

Polling Relay

Can the receiver pull when ready?

A receiver-controlled retrieval model for private or intermittent applications.

Explore Polling Relay

Engineering guides

Answer the operational question first.

Short, practical guides cover cron jobs, background workers, health checks, polling and private networks. The product belongs in the explanation only where it is relevant.

Browse the guides
READ FIRSTHeartbeat vs Health Check

A health check asks whether an application is ready. A heartbeat is evidence sent by the application to its monitor.

Understand the distinction

Build from a real contract

Give every important message a pulse.

Start building