If several messages relate to the same outage, add them to one open task, keeping the time and result of each check. Create a new task when a different address or behavior appears, or an error returns after being fixed.
This is a manual team process. It reduces duplicate work, but does not promise to reduce the number of messages from reChecker. You do not have to turn off notifications to stop creating a task for every repeat.
First, find work that is already underway
Assign someone to review each event first and note where it was handed off. Otherwise, two employees may create identical tasks or both may wait for the other. The process for assigning someone is described in the guide to notification recipients.
Before creating a new task, check the exact address, event time, and current result. A similar description is not enough to combine incidents. Errors on the catalog today and on the form a week ago may be unrelated.
If the connection is clear, add the details to the existing task. If it is unclear, record them separately and ask the person responsible to compare. The person reading the notification does not have to immediately prove that there is one underlying technical cause.
Use one example to distinguish a repeat from a new incident
Suppose the catalog was unavailable in the morning. The team has already assigned the problem to a developer:
| What came in | What to do with the task list |
|---|---|
| 10:05, error at /catalog | Start an investigation and create a task |
| 10:10, same address and response | Add the result to the open task |
| 10:20, the address responds again | Check that it has recovered and add the result |
| 11:00, form at /contacts cannot be submitted | Check separately; do not hide it under the catalog issue |
| A week later, /catalog returns an error again | Record a new incident and link to the previous fix |
This is an illustrative example. It shows why recovery and recurrence call for different actions. The first task has a start and a verified outcome; an error returning after the task is closed needs to be investigated again.
Do not discard intermediate results. They can help you compare event times with server logs. Keep the sequence in one task, not just the latest green status.
Set urgency based on the problem, not the number of messages
Five messages about an old issue are not necessarily more important than one confirmed failure to submit a request. Check what is preventing visitors from completing an action now and which pages are affected.
For a current outage, start by manually checking the address. A technical issue that does not prevent visitors from taking action can be reviewed during the next working period. The event date matters: a message may have arrived or been read after the issue was resolved.
Use Website monitoring to review actual data and channel settings. Your urgency rules do not become a built-in automatic task assignment system. For each significant incident, record the person responsible and the next response deadline.
Check why the same group of issues keeps returning
Once a month, review a frequently recurring incident. The cause may be unfinished work, a new version of a page, or a poor choice of required text. Do not change a setting just to make the messages disappear.
For example, the check text included a promotion date. An editor changed it as planned, and the previous text fragment disappeared. First decide which stable text represents the page you want; then update the setting deliberately. How to choose it is described in the article about an important page.
Add a short note to the group: “one address, same failure in the morning, developer task №12”. Ask another team member to reconstruct the history from the note. If they cannot see the time, address, or new symptom, add the missing details. Shorter correspondence should help with the investigation, not hide differences.