The ops queue closes
itself before standup.
Mesa closes routine ops tickets on its own and asks in Slack before it touches anything risky.
Parcelry ops
Today
Synced 02:16:40
p95 latency on checkout-api
Resolved
Disk at 91% on ingest-worker-3
Resolved
Rotate Stripe webhook secret
Needs approval
Payments on-call gap, Sat night
Resolved
Backup failed, orders-replica-2
Running
14 duplicates from release 4.12
Resolved
2,040 ops teams run Mesa, including
Parcelry
loopwire
Tallgrass Labs
Fernline
Kiteline
Oakhurst Freight
A checkout latency alert at Parcelry in March 2026. Same alert, same runbook, same on-call engineer. The only change was who did the steps.
Before: by hand
Three people awake
PagerDuty pages Ana Ruiz
Ana acknowledges from her phone
Laptop open, VPN, Datadog dashboard
Finds the migration lock on orders
Asks in #payments whether to end the query
Lock released, latency recovers
Postmortem started after standup
From page to recovery
77 minutes
After: with Mesa
One tap from Ana
Alert grouped with 3 related alerts
Matched playbook PB-038, Release table lock
Approval sent to Ana Ruiz in Slack
Approved by Ana Ruiz in Slack
Ended blocking query, p95 at 212 ms
Resolved, postmortem PAY-1182 drafted in Linear
From alert to resolved
1 min 55 s
Closed without a human, August 2026
412,900 tickets across 2,040 workspaces
73%
Counted from resolved tickets with zero human actions after the alert.
Playbook
PB-038
Release table lock
Owner: Payments on-call. Ran 41 times in 2026, approved 41 times.
1
When p95 on checkout-api stays above 800 ms for 3 minutes
auto
2
Look for locks on the orders table older than 30 seconds
auto
3
Ask the on-call engineer in #payments-oncall, query attached
ask
4
On approval, end the blocking query
ask
5
Watch p95 for 30 seconds, then resolve and draft the postmortem
auto
Last run resolved in 1m 55s
02:16:02
From first login
to first closed ticket
in one afternoon.
1
Connect your tools
Sign in to Slack, PagerDuty, Datadog and Linear with OAuth. Mesa reads 90 days of history to learn how your team already works.
About 12 minutes
2
Teach a playbook
Write the runbook in plain English, or let Mesa draft one from your last ten incidents. You edit it like a document.
38 starter playbooks included
3
Choose what asks first
Mark each step auto-run or ask-first. Anything that writes to production asks, in Slack, with the query or diff attached.
Approvals answered in 11 seconds on average
4
Read Monday’s report
What broke, what Mesa closed, what it escalated and what to fix for good, on one page your leadership reads.
Sent Mondays at 8 am in your time zone
Your on-call already has these tools open.
40+ native integrations with one-click OAuth, plus a REST API and webhooks for everything else.
“We retired two on-call rotations. Nobody gets paged for a disk filling up anymore, and the postmortem is drafted before anyone logs in.”
Marcus Feld, CTO at Tallgrass Labs
Pages to humans down 64% in 90 days
“Alerts are triaged before anyone opens Slack. Our on-call week went from dreaded to boring.”
Priya Natarajan, Head of Platform at Loopwire
“The Monday report is the first ops document our leadership reads end to end.”
Sofia Ribeiro, VP Engineering at Fernline
Three plans.
Priced per workspace.
Growth and Scale start with a 14-day trial, no credit card.
Questions.
Anything else? Email hello@mesahq.app and a person on the team replies within one business day.
How does Mesa connect to our tools?
Can Mesa act without asking?
Where is our data stored?
How long does setup take?
Is there a free plan?
What if Mesa gets something wrong?
Connect two tools and turn on a starter playbook. Most teams close their first ticket the same afternoon.
Mesa Labs, Inc. is SOC 2 Type II audited. Your data stays in the US or EU region you choose.