Who this is for. Whoever looks after a live site: the agency developer on support, or the client's own technical owner. Afterwards you can go from a vague report to the log that can answer it.
Finding the right log is most of diagnosing a live problem, because each log answers one kind of question and is silent on the others. A job that did not send an email is in its job runs, not in the template errors; a push that did not deploy is in the deploy log, not anywhere a page would show it.
The logs
| Log | Where | Answers |
|---|---|---|
| Deploy log | The environment's VCS logs in the admin (/cdna-admin/sites/vcs-logs/{environment id}), for super administrators | Did my push deploy (pending, success or failed)? Which commit, branch, message and author? |
| Template Errors | /cdna-admin/sites/errors, per site, from the last 5 minutes to the last 14 days | Which error, raised while rendering which template, on which URL? (Reading the error log) |
| Job runs | Each job's runs, under Customization › Orchestration | Did the job run? Which action failed, with what message? |
The Mailer's batches and the emails in each (/cdna-admin/mailer/batches) | Was the email sent, to whom, and is it waiting, queued, sent, delivered, bounced or failed? | |
| Background tasks | The admin's Jobs screen in the Admin module (/cdna-admin/admin/jobs): the platform's own queue, not Orchestration's jobs | Is a sitemap build, cache clear, index batch, export or job run waiting, running, finished or in error? |
| Audit log | /cdna-admin/admin/audit-log | Who created, changed or deleted a page, post, layout, input type, form, job, menu, redirect, product, order, user or site record, and when? Settings and site data are not in it |
| Notifications | The admin's notifications | What does the platform want an administrator to act on, such as "The master search index needs to be rebuilt"? |
| Billable Requests | /cdna-admin/sites/requests | How many requests did each environment receive per day? Counts only, not the requests themselves |
| Payment provider | The provider's own dashboard, in live mode | Was the money taken? Were its webhooks delivered? |
| Server logs | Core dna, on request | Errors outside templates, and every request, including those the page cache answered |
| Analytics | The site's analytics | What did visitors see: not-found pages, drop-offs, a page that stopped getting traffic? |
A development environment shows an error on the page itself, with its stack trace. Production shows the site's error page and records the error, which is why production problems are read from the logs and reproduced on a development environment.
Which log first
- "The site is down": load it yourself, signed out. If it renders, the report is about one page or one visitor. If not, the server logs, through Core dna.
- "A page shows an error": Template Errors, for the site, over the last hour.
- "My change is not live": the deploy log, then the page's cache headers (Diagnose "my change is not showing up").
- "An email never came": the job runs of the job that sends it, then the email's status in the Mailer batches.
- "Search is out of date": the notifications, then the background tasks.
- "Someone changed something": the audit log; for a setting or site data, ask the administrators.
- "Orders stopped": the payment provider's dashboard and the checkout itself, then Template Errors.
How to find one visitor's failing request among a day's errors, and what an entry records, is on Reading the error log and tracing a request.