Limits & Quotas

• Active resources — enabled HTTP monitors in the workspace right now. Paused monitors don’t count.
• Monthly usage — log bytes, browser events, trace spans and AI analyses accepted since the start of the UTC calendar month. Counters reset at 00:00 UTC on the 1st; unused allowance doesn’t roll over.
• Billing → Usage shows both, with the current period’s start and end, and warns at 80% and 100%.

• HTTP monitors — Free: 3 active, checked at most every 10 minutes; Pro: 10 active, at most every 5 minutes. Projects only organize monitors; they never add monitor capacity.
• Daily SSL & DNS checks — at most once every 24 hours for the hostname saved on each active HTTP monitor. SSL warns at 14, 7 and 1 days before expiry; DNS compares A, AAAA and CNAME records. They don’t count as monitors.
• History — Free keeps 1 day of individual check results and 7 days of daily summaries; Pro keeps 7 days and 90 days.
• Outage and recovery emails are included on both plans. Free alert emails share a limit of 5 per hour and 20 per day per workspace owner.
• Free monitoring capacity is limited during launch. When it is full, a new Free monitor is saved but doesn’t run until a spot opens.
• Pro only: public status pages, incidents, impact analysis on request, and Slack / Teams / Discord / webhook alert channels.

• Logs are measured as canonical validated event JSON bytes before storage. The SDK caps a batch at 500 events and 64 KiB per event — see Ingestion.
• Browser events count every accepted SDK event (page views, vitals, errors, custom events).
• Trace spans are capped by count and by span data: about 1 KB of canonical span JSON per included span (≈ 50 MB on Free, ≈ 1 GB on Pro). Whichever runs out first stops ingest.
• Retention is 1 day on Free and 7 days on Pro for logs, browser data and traces. Add-ons don’t change retention.

Free includes 2 and Pro 20 model-written analyses per UTC month. Each analysis has a fixed ceiling on evidence size, output length and model cost; if a request would exceed it, Mydle shows the evidence summary instead of calling the model. These ceilings are operational safeguards, not published guarantees.

• Logs, browser events or spans: new data is rejected with 429 until the month resets or capacity is added. Data already stored stays searchable for its retention period.
• AI analyses: no new analyses until the month resets or capacity is added.
• Monitors: new monitors can’t be started; existing monitors keep running.
• Over-quota responses share one body: { "code": "QUOTA_EXCEEDED", "meter", "limit", "resetsAt" } plus a Retry-After header.

Pro workspaces can add monthly capacity: +5 HTTP monitors, +1 GiB logs, +50,000 browser events, +1,000,000 spans (with the matching span data), and +20 AI analyses. Add-ons renew with the Pro plan, don’t roll over and don’t change retention or check interval. Increases apply after payment; reductions take effect at the end of the billing period. Free workspaces upgrade to Pro first. See Billing.

Pro, annual and Team subscriptions bought before the current package keep their original limits and price until they’re cancelled. New purchases are monthly Pro only.

• Billing
• Retention
• Log retention
• Browser monitoring
• Projects & Workspaces