Skip to content
LastPing
Start free

Free for individuals

For everything that fails by going quiet.

The AI agent that stopped mid-task, the cron that ran but didn't finish, the pipeline that hung at the last step. Your job pings in; if it goes silent, LastPing opens an incident and pages you before your users notice.

Start monitoring free Follow the project

Free for individuals, fully hosted, first alert in under two minutes.

The LastPing console's Home: the set-up checklist and the monitors that need attention, each with its next action. Demo data. The LastPing console's Home: the set-up checklist and the monitors that need attention, each with its next action. Demo data.

Three monitor types, one engine

Each one uses the same durable detection engine and alert pipeline. AI agents report through the same pings.

Heartbeat and cron

Your job calls a URL when it finishes. Silence opens an incident.

CI/CD signals

A signed webhook from GitHub Actions, GitLab CI or Jenkins.

HTTP uptime

LastPing probes your URL for status code, latency and a keyword.

AI agents, on heartbeats

Start, success, fail and blocked, plus progress steps within a run.

An agent can set up its own monitoring

Point one at mcp.lastping.dev and it creates its own monitor, instruments itself and reports as it works. A hosted MCP server with 50 tools, plus a standalone binary at full tool parity.

  • It calls register_agent and create_monitor, no console visit required.
  • Blocked: it stopped mid-task waiting on a person, and you hear about it.
  • Silence between runs: a silence floor (expect_every_s) catches an agent that stops reporting altogether.
  • Looping or running away: a metric guard caps a number it reports about itself.
  • Finished, achieved nothing: an output assertion fails the run anyway.
How LastPing monitors AI agents
A failed AI agent run in the LastPing console: two of its three declared expectations met and one not met, its duration, and the trace's spans with tokens and estimated API cost. Demo data. A failed AI agent run in the LastPing console: two of its three declared expectations met and one not met, its duration, and the trace's spans with tokens and estimated API cost. Demo data.

Heartbeat and cron: a dead man's switch

Add one curl to the end of your job, set the expected period and a grace window. No library, any language.

  • up, late, down on durable database deadlines: a restart or redeploy never misses a due check.
  • Optional run budget (max_runtime_s) and progress budget (step_timeout_s) for long multi-stage jobs.
  • Exited zero having done nothing? An output assertion fails the run and opens an incident.
  • On recovery the incident closes and a recovery alert fires.
# end of your cron job
curl -fsS https://ping.lastping.dev/<id>
RID=$(uuidgen)
curl -fsS "https://ping.lastping.dev/<id>/start?rid=$RID"
curl -fsS "https://ping.lastping.dev/<id>/step?rid=$RID&step=extract"
curl -fsS "https://ping.lastping.dev/<id>?rid=$RID"

CI/CD: pipeline visibility via signed webhooks

GitHub Actions, GitLab CI and Jenkins (webhook plugin) send a signed webhook when a job starts and when it finishes. Zero YAML config on the LastPing side.

  • Alerts on failure, on silence (started, never reported back), on consecutive-failure streaks, and on recovery.
  • Payloads are HMAC-signed; LastPing verifies before processing.
  • The alert deep-links to the exact run: one click to the log.
Monitor GitHub Actions
The LastPing console's Incidents list: each incident with its monitor, its cause, how long it lasted and how many alerts went out, plus the median time to resolve. Demo data. The LastPing console's Incidents list: each incident with its monitor, its cause, how long it lasted and how many alerts went out, plus the median time to resolve. Demo data.

HTTP uptime: LastPing probes your URL

Give LastPing a URL, no agent to install on your server. It checks status code, latency and keyword presence on your interval.

  • Catches non-2xx or timeout, latency spikes, a missing keyword, TLS failures and DNS failures.
  • SSRF-guarded: probes never reach private IP ranges or cloud-metadata endpoints.
HTTP checks
The LastPing console's Monitors list: state, schedule, last ping, next due, the last 20 runs, uptime and alert routing for each monitor. Demo data. The LastPing console's Monitors list: state, schedule, last ping, next due, the last 20 runs, uptime and alert routing for each monitor. Demo data.

Route the right signal to the right channel

Nine delivery channels, a routing matrix per monitor, custom message templates, and flap damping so you only get paged when it matters.

  • email
  • Slack
  • Discord
  • Telegram
  • webhook
  • Microsoft Teams
  • Google Chat
  • Pushover
  • ntfy
  • Email is the default: it fires on the very first check, no config needed.
  • Templates take variables such as {check_name}, {status} and {run_url}, with a live preview, per event type.
  • A blip that recovers within a short window is held and cancelled. Per-channel rate caps suppress duplicates.
Example alert routing matrix, per destination and per event
ChannelDownFailBlockedRecover
emailyesyesyesno
slackyesyesyesyes
telegramyesnoyesno
pushoveryesnonono
webhookyesyesyesyes

Illustrative. Every one of the nine destinations gets its own row. Blocked fires when an agent stops mid-task waiting on a person. started, success, every-run and note are routable too, never by default.

Four ways in, each for what it is good at.

Monitors, destinations, routing and status pages are on all four.

Web console

Home shows what needs you now. Light and dark themes, and a command palette.

REST API

A published OpenAPI document, partial PATCH, and a Prometheus /metrics endpoint per project.

MCP

What an agent would do or ask, from creating a monitor to reading a failed run.

Terraform

Declarative configuration in version control: lastping-dev/lastping.

Public status pages, no extra service needed

Backed by the same monitors that already power your alerts, with no duplicate configuration.

  • A public status URL per project, with no separate hosting and no additional cost.
  • A live, embeddable badge for any README or docs page.
  • You choose which monitors show; internal cron jobs can stay private.
See LastPing's own status page
LastPing's public status page as captured on 3 October 2026: 'All systems operational' above six component checks (ping ingest, App/API health, the remote MCP server, the app login page, API docs and the marketing site), each with an uptime bar covering up to 90 days. All six carry real incident marks, left visible rather than hidden. LastPing's public status page as captured on 3 October 2026: 'All systems operational' above six component checks (ping ingest, App/API health, the remote MCP server, the app login page, API docs and the marketing site), each with an uptime bar covering up to 90 days. All six carry real incident marks, left visible rather than hidden.
Snapshot taken 3 October 2026.

Start monitoring in minutes

Free for individuals. One URL to ping, or let your AI assistant set it up.

Start free Connect your AI assistant