Skip to main content

Uptime

The Uptime module is the outside view: Wakora's points of presence probe your public sites over HTTPS and report reachability, status and latency per site, per vantage. This closes the one gap the agent cannot see from inside the box: a healthy server whose ingress is dead. The mechanics and the honesty rules are described in Availability.

What gets probed​

Your genuinely public sites, automatically: names discovered on your hosts that resolve to public addresses. Internal names, IP literals and catch-alls are excluded, so a split-horizon lab name never produces a false "down from everywhere". There is nothing to configure; the target list follows your fleet.

Synthetic probing is included on every plan, up to the plan's sites quota. On a workspace without a card on file, sites beyond the quota are still listed, under "beyond the plan quota", but not probed from our vantage points until you upgrade or add a card; the agent's own checks on the host are unaffected. With a card on file every enrolled site is probed and the surplus bills as site overage.

Reading the page​

Each site shows its per-vantage verdicts and response times, with uptime bars over time. A DNS hiccup at one vantage reads as unknown, never as down; a genuinely dead domain is the DNS liveness signal on the host's Sites block, not an uptime verdict. A site served by several of your hosts (a cluster behind one name) is one target: the row names the first host and counts the others ("+2", the full list in the tooltip), and the outage rule reads the site as fine from the inside when it is fine on any one of them, so a node you took out of rotation does not turn a healthy site into an ingress outage.

Transactions​

Beyond single-page checks, the Transactions tab runs multi-step flows from the same vantages: up to ten HTTP steps with variables, value extraction between steps, cookies, and assertions on status, body content and timing. Think "open the pricing page, add to cart, reach checkout": the flow fails on the first broken step and tells you which one.

The editor has a test run button that executes the flow immediately and shows the step-by-step trace before you save. Flows needing credentials use test accounts you provide for exactly this purpose (stored encrypted; deliberately not the host-secret machinery, since these are credentials you issue to be used from outside). Flows run every few minutes from each vantage; a flow failing from all vantages raises the transaction rule, which ships disabled until you arm it. Each flow counts as one monitored site. Available on Pro and above.

Alerting on it​

The ingress outage rule (all vantages failing while the inside view is healthy) is armed automatically on workspaces where synthetic probes flow, at warning severity, after a long clean burn-in of the detector. Raising it to critical (and the phone call that implies) stays your explicit choice in Alerting. The transaction failure rule stays opt-in.

Allow-listing the probes​

The vantage addresses are fixed and published (with the probe user agent) on the Wakora website's synthetics page, so a WAF or rate limiter can be taught to let them through rather than fight them.

E-mail deliverability​

The module's second tab watches whether mail from your domains actually lands. For every sender domain in your fleet (discovered sites that publish MX records), Wakora's infrastructure checks, every few hours:

  • SPF: the record exists, and it stays under the 10-lookup limit that silently breaks SPF.
  • DMARC: a policy is published.
  • DKIM: signing keys are discoverable under the common selectors.
  • Blocklists: whether the host's public address or its MX addresses sit on the major DNSBLs. An unreachable or rate-limited blocklist reads as unknown, never as listed.

Each domain gets a score and the row tells you exactly which piece is missing. A blocklist-listing alert rule exists but ships disabled (some lists are noisy by nature); arm it per workspace when you want it. The agent's own mail-server probes (queues, bounce rates) stay on the host side, see the service catalog.