An outage spotted in thirty seconds

Every plan now checks more often: a full pass every 5 minutes on Discover, every 2 minutes on Pro, every minute on Pro Plus. But a payment page that goes down right after a pass stays down until the next one before PostShip sees it: enough to lose an order. On Pro Plus, that delay drops to thirty seconds.

Closer passes, on every plan

The full pass is the one that requests every page with its files, reads the HTML, replays the journeys and reads the heartbeats. It now comes back every 5 minutes on Discover instead of fifteen, every 2 minutes on Pro instead of five, and every minute on Pro Plus instead of two. Nothing to set: the cadence follows the workspace's plan, and the “Check” line on Billing shows it.

A light probe, every thirty seconds

Between two passes, every page watched on Pro Plus gets a light request every thirty seconds: the status, the response time, the expected content. As long as the answer says the same thing as the last pass, nothing is written and nothing goes out. Your history doesn't fill up with identical passes.

At the slightest change, the full pass

As soon as a page changes state — it goes down, or it comes back —, the probe wakes the engine for that project. The engine is what confirms the outage according to your rules, records it, opens the incident and warns you, exactly as usual: the probe never sends an alert on its own. The return to green follows the same path, and “Recovered” arrives sooner too.

A flickering site doesn't cause a burst: a project that was just woken isn't woken again for a minute.

What doesn't change

• The probe only watches pages: the sitemap, the certificate, heartbeats and forms follow the cycle.
• Discover and Pro have no probe: an outage is seen at the next pass, at most 5 or 2 minutes later.
• Every deploy is still checked the second the host announces it, on every plan.