Monitoring log and tuning alerts
What the downloadable monitoring log records, how to read it, and how to use it to judge whether alert levels suit your store.
On this page
PRO’s alert levels are starting values. They were chosen from how stores usually behave, not from measurements of your store, so they may be too sensitive or too relaxed for you. The monitoring log gives you the numbers to judge that.
What it is
Every time a monitor evaluates something, it writes one row: when, which monitor, what it looked at, the figures it used, and its verdict. This includes checks that found nothing, so you can see how close a quiet day came to raising an alert.
- Open Commerce Health → PRO license.
- In Monitoring Log, click Download the log (CSV).
The log is kept for 30 days, then older rows are removed automatically. It stays in your own database and is never sent to Kodenzo. It holds no customer details: only counts, percentages, gateway or plugin names, and metric values.
What each monitor logs
| Monitor | One row per… | The figures |
|---|---|---|
| Stuck Payments | enabled gateway, about every 10 minutes (even with no orders) | Unfinished orders and orders in the last hour, the share, the usual share and how many orders it is based on, and orders, paid and failed over 24 hours |
| Payment Gateway | gateway with checkout activity in the last 60 minutes, every 5 minutes | Attempts, failures, failure percentage, average time in milliseconds |
| Core Web Vitals | metric (LCP, INP or CLS) with at least 3 visits in 24 hours, every 5 minutes | The 75th percentile, the number of visits |
| Plugin Conflict | plugin change, once, when it is judged | The kind of change, new JavaScript errors, payment failures before and after |
The verdict column says “none” when nothing was raised, or the level that was raised (low, medium, high, critical). For Core Web Vitals, “critical” in the log means the metric was rated Poor.
Not logged: JavaScript errors (their issues already carry the counts) and the shipping, coupon and theme checks.
Using it
- Too many alerts? Look at the rows with a verdict and the figures behind them. If a normal day sits just over a line, that line is too low for your store.
- A problem you knew about but got no alert? Find the rows around that time. If the figures were close but the verdict is “none”, the line is too high.
- Compare days and hours. A store with few orders at night behaves differently from one at lunchtime.
The levels themselves can’t be changed in settings in this version. If you want help judging them, you can send the CSV to support yourself.
Limits
- A row exists only when a monitor had something to evaluate. A gateway with no checkout activity has no Payment Gateway row.
- With several gateways, Stuck Payments alone writes about 144 rows a day for each one, and Payment Gateway up to 288. That is why the log is capped at 30 days.
- The log shows what the monitors saw. It can’t show a problem they can’t see. See What PRO can and cannot detect.
All the default values are listed in the PRO thresholds reference.
Still need help?
Search the docs first — or open a ticket and our team will pick it up from the Support screen.