Review automations that failed or double-ran
A weekly log of failed and duplicate automation runs, with the time spent fixing them, so problems surface before customers find them.
IT and admin · Ecommerce · On a schedule · Updated
Slack
Datadog
Sentry
New Relic
Grafana
Linear
Asana
ClickUp
Jira
TrelloHow do you review failed or double-run automations?
To review failed or double-run automations, set a weekly Loop in Maito. It reads your logs, lists the runs that failed or ran twice, estimates the time spent fixing them and posts the list in Slack. It can open tickets in your tracker for the ones that need work.
How it's done today
Failures are found by customers, and nobody knows how much time fixing them costs.
What Maito does
- Read the week's run logs
- List failures and duplicates
- Queue recoverable cases
- Add up correction time against the benefit
Example
Loop run: Friday 4:00 PM · Automation log review
This week: 1,284 runs, 6 failed, 2 duplicates. • Shipping label sync: 4 failures, all Tuesday, the carrier API timed out. 3 recovered by retry; order 10431 still has no label. • Order confirmation email: sent twice to Lucy Bell and Dee Harmon on orders 10440 and 10447. • Low stock alert: 2 failures, a missing SKU field. Catch: the double emails began after the Oct 14 change to the order webhook, and 31 others got a second email earlier this month. That's more than the 2 flagged in this week's log. Source: automation logs. Estimated 3.5 hours fixing this week, against about 9 hours saved.
@Maito put order 10431 in the tracker as urgent.
Added as urgent: order 10431, no shipping label, assigned to you. I included the log lines from Tuesday 2:14 PM.
Maito (4:00 PM):
This week: 1,284 runs, 6 failed, 2 duplicates.
• Shipping label sync: 4 failures, all Tuesday, the carrier API timed out. 3 recovered by retry; order 10431 still has no label.
• Order confirmation email: sent twice to Lucy Bell and Dee Harmon on orders 10440 and 10447.
• Low stock alert: 2 failures, a missing SKU field.
Catch: the double emails began after the Oct 14 change to the order webhook, and 31 others got a second email earlier this month. That's more than the 2 flagged in this week's log. Source: automation logs.
Estimated 3.5 hours fixing this week, against about 9 hours saved.
Ben (4:12 PM):
put order 10431 in the tracker as urgent.
Maito (4:13 PM):
Added as urgent: order 10431, no shipping label, assigned to you. I included the log lines from Tuesday 2:14 PM.
How to set it up
- Ask your agent
“Every Friday at 4, read this week's automation logs and post failures, duplicates and recoverable cases in #ops-admin, with an estimate of time spent fixing.”
- Connect logs
- Datadog
- Sentry
- New Relic
- Grafana
- Connect task tracker
- Linear
- Asana
- ClickUp
- Jira
- Trello
Not connected yet? Your agent sends you a link to sign in.
What it won't do
It sees only what your logs record. An automation that fails silently without logging an error won't show up.
Questions
Datadog, Sentry, New Relic or Grafana for logs, and Linear, Asana, ClickUp, Jira or Trello for tasks.
No. It reports and can open tasks. A person fixes them.
Yes. It can run as a Loop on any schedule, or check every few minutes.
