Data retention
PathWatch keeps several kinds of data, and each kind has its own retention window. Your plan sets five numbers: how long full-resolution check results live, how long summary aggregates extend beyond them, the total history window the two add up to, how long browser artifacts (screenshots) stick around, and how long logs are kept.
Retention at a glance
| Data | Free | Starter | Pro | Business |
|---|---|---|---|---|
| Full-resolution check results | 7 days | 7 days | 14 days | 30 days |
| Summary aggregates (beyond raw) | +83 days | +173 days | +351 days | +700 days |
| Total history window | 90 days | 180 days | 365 days | 730 days |
| Browser artifacts (screenshots) | — | — | 1 day | 90 days |
| Logs (audit log & alert history) | 30 days | 90 days | 180 days | 365 days |
The first two rows always add up to the total: e.g. Free keeps every individual check result for 7 days, then daily summaries carry the history out to 90 days total.
How the rollup model works
Every check produces a raw result — the individual row with its exact timestamp, status, response time, status code, error message, region and check-type metadata. Raw results power the detailed views: per-check drill-downs, error messages, region comparisons, and the fine-grained (5-minute / hourly) response time graphs.
Once a raw result ages past your plan’s full-resolution window, a periodic sweep rolls it up into a daily summary bucket and deletes the raw row. Each daily bucket records, per monitor:
- total checks, and counts of successful / degraded / error / timeout / skipped checks
- uptime percentage for the day
- average, minimum and maximum response time
As summaries age further they consolidate into coarser buckets (weekly, then monthly), so long-range trend charts, uptime percentages and SLA reports keep working across your full plan window without storing every individual result.
What disappears when
| Age of data | What you can still see |
|---|---|
| Within the full-resolution window | Everything: individual results, status codes, error messages, per-region detail, 5-minute and hourly graphs |
| Past full-resolution, within the total window | Daily (and later coarser) summaries: uptime %, check counts by status, avg/min/max response times. Individual results and their error details are gone |
| Past the total window | Data is no longer retained under your plan |
Two practical consequences:
- Debug soon after an incident. The exact error message, status code and region breakdown for a specific failed check are only available while the raw result exists (7–30 days depending on plan).
- Trends survive. Uptime and response-time history remain available for the full window — they just lose per-check granularity.
Retention is enforced by a periodic background sweep, so data may linger slightly past its window before being rolled up or removed.
Browser artifacts (screenshots)
Browser Lite and
Browser Flow checks can capture
screenshots (e.g. screenshot_on_failure, explicit screenshot
steps). These artifacts have their own, shorter window —
separate from the check result they belong to:
| Plan | Artifact retention |
|---|---|
| Free | Not retained |
| Starter | Not retained |
| Pro | 1 day |
| Business | 90 days |
On Pro, screenshots are there to debug a failure that just happened — view or save them within a day. On Business they persist for 90 days. After the artifact window the screenshot is deleted even if the check result itself is still within the full-resolution window.
Logs
The audit log (who changed what, when) and the alert history (every notification PathWatch sent, to which channel, with what outcome) are both kept for the plan’s log window: 30 / 90 / 180 / 365 days. Entries older than that are permanently deleted by the sweep.
Keeping raw data longer
If you need individual check results beyond your plan’s full-resolution window, export them before they roll up:
- Per monitor: monitor detail page → Export → CSV / JSON,
or
GET /api/v1/monitors/:id/checks/exportvia the REST API. - Account-wide: see Data export & GDPR.
Exports are point-in-time snapshots — schedule them on a cadence shorter than your raw window if you want a continuous external archive.
You can read your org’s current total window programmatically via
GET /api/v1/retention, which returns the retention period your
active plan grants.
Plan changes
- Upgrading extends the windows from that point forward. It does not resurrect data that was already rolled up or deleted — a raw result that became a daily summary stays a summary.
- Downgrading shrinks the windows. On the next sweep, data older than the new plan’s windows is rolled up or removed. Export anything you want to keep before downgrading.
Plan numbers on this page reflect the current packages — see Limits & quotas for the live, build-time plan comparison and Billing for how plan switches take effect.