A blameless postmortem, in ten parts.
Ten parts that help a team learn from a bad day without looking for someone to blame. Fits a 45-minute review.
Running the 45 minutes
- The day before: the author fills parts 01–04 and shares them.
- First 30 minutes: walk the timeline together and fill parts 05–07.
- Last 15 minutes: agree on the action items, one owner each.
The example is Acme Shop, a made-up online store. Every section uses the same story.
On this page
-
Header
The facts a reader needs before anything else.
Fields
- Title
- Date
- Severity
- Status
- Authors
- Tickets
Example
- Title
- Checkout down for 47 minutes
- Date
- 4 Mar 2026, 14:02 to 14:49 UTC
- Severity
- SEV1 (all customers could not pay)
- Status
- Final, reviewed 7 Mar
- Authors
- Priya (incident lead), Sam
- Tickets
- OPS-412 to OPS-416, SHOP-88
-
01 Summary
Three sentences: what broke, who felt it, how it ended.
Hints
- Write it last, read it first.
- No jargon a new hire would need to look up.
Example
For 47 minutes nobody could pay at Acme Shop. The certificate on our payments API had expired, so checkout could not reach it. We renewed it by hand and checkout came back at 14:49 UTC.
-
02 Impact
Who was affected, how badly, and for how long. Use numbers.
Hints
- Count people or orders, not just minutes.
- Say what customers saw, in their words.
Example
- All checkouts failed for 47 min: about 1,900 orders not placed.
- Customers saw "Payment declined", so many thought their card was the problem.
- No data lost. No one was charged twice.
-
03 Detection
How we found out, and how long that took.
Hints
- Write the time from "it broke" to "a person knew".
- Did an alert fire? If a customer told you first, say so.
Example
No alert fired. Support saw angry posts on social media at 14:21, 19 minutes after checkout broke. Our health checks called the API from inside our network, behind the load balancer where TLS had already ended, so they never saw the expired certificate.
An outside check would have caught this in 1 minute, not 19. That's the kind of check we build at StatusTick.
See how -
04 Timeline
Times in UTC. What we saw, what we thought, what we did.
Hints
- Include the wrong guesses. They show where the tools misled you.
- Copy times from chat and alerts, not from memory.
Example
- 14:02Certificate expires. Checkout starts failing.
- 14:21Support sees posts about failed payments and pages on-call.
- 14:28On-call checks the card provider first, because the error said "declined".
- 14:31Status page updated: "Payments are failing. We're working on it."
- 14:39Expired certificate found. The runbook still points to the old DNS provider.
- 14:49Certificate renewed by hand. Checkout works again.
-
05 Contributing factors
The conditions that made this possible, and the ones that made it last longer.
Hints
- Ask "why did the system allow this?", never "who did this?".
- Keep two chains apart: why it broke, and why it took so long.
Example
Why did it break?
- A1The payments API certificate expired.
- A2Auto-renewal had been failing since we moved DNS providers in January.
- A3The renewal job still used the old provider's API key.
Why did it take 47 minutes?
- B1No alert on renewal errors or on soon-to-expire certificates.
- B2Our checks called the API behind the load balancer, where TLS had already ended, so they never saw the expired certificate.
- B3The error message said "declined" and sent us to the card provider first.
- B4The runbook described the old DNS setup.
-
06 What went well
The things that helped. There are always some.
Hints
- Write this part while the good is fresh, it fades first.
- Name tools and habits worth keeping.
- Also write down where you got lucky. Luck is not a plan.
Example
- Support paged on-call within minutes of the first posts.
- No double charges: the payment retries were safe to repeat.
- The status page was updated within 10 minutes of the page.
-
07 What was hard
Where people struggled: tools, information, timing.
Hints
- Describe the situation, not anyone's skill.
- "We didn't know X" is more useful than "X was missed".
Example
- The error message pointed at the wrong system.
- The renewal steps in the runbook were 8 months old.
- It happened during the busiest hour of the week.
-
08 Action items
Each one fixes a factor from part 05. Owner, date and ticket for every row.
Hints
- One owner per row, not a team.
- Mark each as prevent, detect or recover, and check you have some of each.
Example
Action Type Fixes Owner Due Ticket Fix auto-renewal for the new DNS provider and test a renewal in staging Prevent A2, A3 Priya 14 Mar OPS-412 Add a "What depends on this provider?" step to the migration checklist Prevent A2 Priya 28 Mar OPS-416 Alert on renewal errors and on certificates expiring within 14 days Detect A1, B1 Sam 10 Mar OPS-413 Check checkout from outside our network every minute Detect B2 Sam 14 Mar OPS-415 Show "Temporary problem, you were not charged" instead of "Payment declined" when the payments API is down Recover B3 Lee 21 Mar SHOP-88 Update the certificate runbook for the new DNS setup, with a 5-minute manual renewal Recover B4 Jo 17 Mar OPS-414 -
09 Thanks
Who helped, and how.
Hints
- Thank the people who noticed and spoke up, not only the fixers.
Example
Thanks to Lee in support for paging fast, to Priya for staying calm on the call, and to Jo for writing the renewal steps down while fixing them.
-
10 Links
Dashboards, chat logs, graphs and tickets.
Hints
- Link the incident chat so readers can see the real order of events.
Example
- Incident chat: #inc-2026-03-04-checkout
- Certificate dashboard, payments API error graph
That's all ten parts.
Take it to your next review.
Blameless postmortem
From healthcheck.wtf. Free to use, change and share.
Ten parts that help a team learn from a bad day without looking for someone to blame. Fits a 45-minute review.
Running the 45 minutes
- The day before: the author fills parts 01–04 and shares them.
- First 30 minutes: walk the timeline together and fill parts 05–07.
- Last 15 minutes: agree on the action items, one owner each.
Header
The facts a reader needs before anything else.
Title:
Date:
Severity:
Status:
Authors:
Tickets:
01 · Summary
Three sentences: what broke, who felt it, how it ended.
Hints
- Write it last, read it first.
- No jargon a new hire would need to look up.
Write yours here.
02 · Impact
Who was affected, how badly, and for how long. Use numbers.
Hints
- Count people or orders, not just minutes.
- Say what customers saw, in their words.
Write yours here.
03 · Detection
How we found out, and how long that took.
Hints
- Write the time from "it broke" to "a person knew".
- Did an alert fire? If a customer told you first, say so.
Write yours here.
04 · Timeline
Times in UTC. What we saw, what we thought, what we did.
Hints
- Include the wrong guesses. They show where the tools misled you.
- Copy times from chat and alerts, not from memory.
Write yours here.
05 · Contributing factors
The conditions that made this possible, and the ones that made it last longer.
Hints
- Ask "why did the system allow this?", never "who did this?".
- Keep two chains apart: why it broke, and why it took so long.
Write yours here.
06 · What went well
The things that helped. There are always some.
Hints
- Write this part while the good is fresh, it fades first.
- Name tools and habits worth keeping.
- Also write down where you got lucky. Luck is not a plan.
Write yours here.
07 · What was hard
Where people struggled: tools, information, timing.
Hints
- Describe the situation, not anyone's skill.
- "We didn't know X" is more useful than "X was missed".
Write yours here.
08 · Action items
Each one fixes a factor from part 05. Owner, date and ticket for every row.
Hints
- One owner per row, not a team.
- Mark each as prevent, detect or recover, and check you have some of each.
| Action | Type | Fixes | Owner | Due | Ticket |
|---|---|---|---|---|---|
09 · Thanks
Who helped, and how.
Hints
- Thank the people who noticed and spoke up, not only the fixers.
Write yours here.
10 · Links
Dashboards, chat logs, graphs and tickets.
Hints
- Link the incident chat so readers can see the real order of events.
Write yours here.
Blameless postmortem
From healthcheck.wtf. Free to use, change and share.
Ten parts that help a team learn from a bad day without looking for someone to blame. Fits a 45-minute review.
Running the 45 minutes
- The day before: the author fills parts 01–04 and shares them.
- First 30 minutes: walk the timeline together and fill parts 05–07.
- Last 15 minutes: agree on the action items, one owner each.
The example is Acme Shop, a made-up online store. Every section uses the same story.
Header
The facts a reader needs before anything else.
Example
Title: Checkout down for 47 minutes
Date: 4 Mar 2026, 14:02 to 14:49 UTC
Severity: SEV1 (all customers could not pay)
Status: Final, reviewed 7 Mar
Authors: Priya (incident lead), Sam
Tickets: OPS-412 to OPS-416, SHOP-88
Title:
Date:
Severity:
Status:
Authors:
Tickets:
01 · Summary
Three sentences: what broke, who felt it, how it ended.
Hints
- Write it last, read it first.
- No jargon a new hire would need to look up.
Example
For 47 minutes nobody could pay at Acme Shop. The certificate on our payments API had expired, so checkout could not reach it. We renewed it by hand and checkout came back at 14:49 UTC.
Write yours here.
02 · Impact
Who was affected, how badly, and for how long. Use numbers.
Hints
- Count people or orders, not just minutes.
- Say what customers saw, in their words.
Example
- All checkouts failed for 47 min: about 1,900 orders not placed.
- Customers saw "Payment declined", so many thought their card was the problem.
- No data lost. No one was charged twice.
Write yours here.
03 · Detection
How we found out, and how long that took.
Hints
- Write the time from "it broke" to "a person knew".
- Did an alert fire? If a customer told you first, say so.
Example
No alert fired. Support saw angry posts on social media at 14:21, 19 minutes after checkout broke. Our health checks called the API from inside our network, behind the load balancer where TLS had already ended, so they never saw the expired certificate.
Write yours here.
04 · Timeline
Times in UTC. What we saw, what we thought, what we did.
Hints
- Include the wrong guesses. They show where the tools misled you.
- Copy times from chat and alerts, not from memory.
Example
- 14:02 Certificate expires. Checkout starts failing.
- 14:21 Support sees posts about failed payments and pages on-call.
- 14:28 On-call checks the card provider first, because the error said "declined".
- 14:31 Status page updated: "Payments are failing. We're working on it."
- 14:39 Expired certificate found. The runbook still points to the old DNS provider.
- 14:49 Certificate renewed by hand. Checkout works again.
Write yours here.
05 · Contributing factors
The conditions that made this possible, and the ones that made it last longer.
Hints
- Ask "why did the system allow this?", never "who did this?".
- Keep two chains apart: why it broke, and why it took so long.
Example
Why did it break?
- A1 The payments API certificate expired.
- A2 Auto-renewal had been failing since we moved DNS providers in January.
- A3 The renewal job still used the old provider's API key.
Why did it take 47 minutes?
- B1 No alert on renewal errors or on soon-to-expire certificates.
- B2 Our checks called the API behind the load balancer, where TLS had already ended, so they never saw the expired certificate.
- B3 The error message said "declined" and sent us to the card provider first.
- B4 The runbook described the old DNS setup.
Write yours here.
06 · What went well
The things that helped. There are always some.
Hints
- Write this part while the good is fresh, it fades first.
- Name tools and habits worth keeping.
- Also write down where you got lucky. Luck is not a plan.
Example
- Support paged on-call within minutes of the first posts.
- No double charges: the payment retries were safe to repeat.
- The status page was updated within 10 minutes of the page.
Write yours here.
07 · What was hard
Where people struggled: tools, information, timing.
Hints
- Describe the situation, not anyone's skill.
- "We didn't know X" is more useful than "X was missed".
Example
- The error message pointed at the wrong system.
- The renewal steps in the runbook were 8 months old.
- It happened during the busiest hour of the week.
Write yours here.
08 · Action items
Each one fixes a factor from part 05. Owner, date and ticket for every row.
Hints
- One owner per row, not a team.
- Mark each as prevent, detect or recover, and check you have some of each.
Example
| Action | Type | Fixes | Owner | Due | Ticket |
|---|---|---|---|---|---|
| Fix auto-renewal for the new DNS provider and test a renewal in staging | Prevent | A2, A3 | Priya | 14 Mar | OPS-412 |
| Add a "What depends on this provider?" step to the migration checklist | Prevent | A2 | Priya | 28 Mar | OPS-416 |
| Alert on renewal errors and on certificates expiring within 14 days | Detect | A1, B1 | Sam | 10 Mar | OPS-413 |
| Check checkout from outside our network every minute | Detect | B2 | Sam | 14 Mar | OPS-415 |
| Show "Temporary problem, you were not charged" instead of "Payment declined" when the payments API is down | Recover | B3 | Lee | 21 Mar | SHOP-88 |
| Update the certificate runbook for the new DNS setup, with a 5-minute manual renewal | Recover | B4 | Jo | 17 Mar | OPS-414 |
| Action | Type | Fixes | Owner | Due | Ticket |
|---|---|---|---|---|---|
09 · Thanks
Who helped, and how.
Hints
- Thank the people who noticed and spoke up, not only the fixers.
Example
Thanks to Lee in support for paging fast, to Priya for staying calm on the call, and to Jo for writing the renewal steps down while fixing them.
Write yours here.
10 · Links
Dashboards, chat logs, graphs and tickets.
Hints
- Link the incident chat so readers can see the real order of events.
Example
- Incident chat: #inc-2026-03-04-checkout
- Certificate dashboard, payments API error graph
Write yours here.
# Blameless postmortem From healthcheck.wtf. Free to use, change and share. Ten parts that help a team learn from a bad day without looking for someone to blame. Fits a 45-minute review. **Running the 45 minutes** 1. **The day before:** the author fills parts 01–04 and shares them. 2. **First 30 minutes:** walk the timeline together and fill parts 05–07. 3. **Last 15 minutes:** agree on the action items, one owner each. ## Header _The facts a reader needs before anything else._ - **Title:** - **Date:** - **Severity:** - **Status:** - **Authors:** - **Tickets:** ## 01 · Summary _Three sentences: what broke, who felt it, how it ended._ Hints: - Write it last, read it first. - No jargon a new hire would need to look up. _Write yours here._ ## 02 · Impact _Who was affected, how badly, and for how long. Use numbers._ Hints: - Count people or orders, not just minutes. - Say what customers saw, in their words. _Write yours here._ ## 03 · Detection _How we found out, and how long that took._ Hints: - Write the time from "it broke" to "a person knew". - Did an alert fire? If a customer told you first, say so. _Write yours here._ ## 04 · Timeline _Times in UTC. What we saw, what we thought, what we did._ Hints: - Include the wrong guesses. They show where the tools misled you. - Copy times from chat and alerts, not from memory. _Write yours here._ ## 05 · Contributing factors _The conditions that made this possible, and the ones that made it last longer._ Hints: - Ask "why did the system allow this?", never "who did this?". - Keep two chains apart: why it broke, and why it took so long. _Write yours here._ ## 06 · What went well _The things that helped. There are always some._ Hints: - Write this part while the good is fresh, it fades first. - Name tools and habits worth keeping. - Also write down where you got lucky. Luck is not a plan. _Write yours here._ ## 07 · What was hard _Where people struggled: tools, information, timing._ Hints: - Describe the situation, not anyone's skill. - "We didn't know X" is more useful than "X was missed". _Write yours here._ ## 08 · Action items _Each one fixes a factor from part 05. Owner, date and ticket for every row._ Hints: - One owner per row, not a team. - Mark each as prevent, detect or recover, and check you have some of each. | Action | Type | Fixes | Owner | Due | Ticket | | --- | --- | --- | --- | --- | --- | | | | | | | | ## 09 · Thanks _Who helped, and how._ Hints: - Thank the people who noticed and spoke up, not only the fixers. _Write yours here._ ## 10 · Links _Dashboards, chat logs, graphs and tickets._ Hints: - Link the incident chat so readers can see the real order of events. _Write yours here._ # Blameless postmortem From healthcheck.wtf. Free to use, change and share. Ten parts that help a team learn from a bad day without looking for someone to blame. Fits a 45-minute review. **Running the 45 minutes** 1. **The day before:** the author fills parts 01–04 and shares them. 2. **First 30 minutes:** walk the timeline together and fill parts 05–07. 3. **Last 15 minutes:** agree on the action items, one owner each. > The example is Acme Shop, a made-up online store. Every section uses the same story. ## Header _The facts a reader needs before anything else._ Example: > - **Title:** Checkout down for 47 minutes > - **Date:** 4 Mar 2026, 14:02 to 14:49 UTC > - **Severity:** SEV1 (all customers could not pay) > - **Status:** Final, reviewed 7 Mar > - **Authors:** Priya (incident lead), Sam > - **Tickets:** OPS-412 to OPS-416, SHOP-88 - **Title:** - **Date:** - **Severity:** - **Status:** - **Authors:** - **Tickets:** ## 01 · Summary _Three sentences: what broke, who felt it, how it ended._ Hints: - Write it last, read it first. - No jargon a new hire would need to look up. Example: > For 47 minutes nobody could pay at Acme Shop. The certificate on our payments API had expired, so checkout could not reach it. We renewed it by hand and checkout came back at 14:49 UTC. _Write yours here._ ## 02 · Impact _Who was affected, how badly, and for how long. Use numbers._ Hints: - Count people or orders, not just minutes. - Say what customers saw, in their words. Example: > - All checkouts failed for 47 min: about 1,900 orders not placed. > - Customers saw "Payment declined", so many thought their card was the problem. > - No data lost. No one was charged twice. _Write yours here._ ## 03 · Detection _How we found out, and how long that took._ Hints: - Write the time from "it broke" to "a person knew". - Did an alert fire? If a customer told you first, say so. Example: > No alert fired. Support saw angry posts on social media at 14:21, 19 minutes after checkout broke. Our health checks called the API from inside our network, behind the load balancer where TLS had already ended, so they never saw the expired certificate. _Write yours here._ ## 04 · Timeline _Times in UTC. What we saw, what we thought, what we did._ Hints: - Include the wrong guesses. They show where the tools misled you. - Copy times from chat and alerts, not from memory. Example: > - **14:02** Certificate expires. Checkout starts failing. > - **14:21** Support sees posts about failed payments and pages on-call. > - **14:28** On-call checks the card provider first, because the error said "declined". > - **14:31** Status page updated: "Payments are failing. We're working on it." > - **14:39** Expired certificate found. The runbook still points to the old DNS provider. > - **14:49** Certificate renewed by hand. Checkout works again. _Write yours here._ ## 05 · Contributing factors _The conditions that made this possible, and the ones that made it last longer._ Hints: - Ask "why did the system allow this?", never "who did this?". - Keep two chains apart: why it broke, and why it took so long. Example: > **Why did it break?** > > - **A1** The payments API certificate expired. > - **A2** Auto-renewal had been failing since we moved DNS providers in January. > - **A3** The renewal job still used the old provider's API key. > > **Why did it take 47 minutes?** > > - **B1** No alert on renewal errors or on soon-to-expire certificates. > - **B2** Our checks called the API behind the load balancer, where TLS had already ended, so they never saw the expired certificate. > - **B3** The error message said "declined" and sent us to the card provider first. > - **B4** The runbook described the old DNS setup. _Write yours here._ ## 06 · What went well _The things that helped. There are always some._ Hints: - Write this part while the good is fresh, it fades first. - Name tools and habits worth keeping. - Also write down where you got lucky. Luck is not a plan. Example: > - Support paged on-call within minutes of the first posts. > - No double charges: the payment retries were safe to repeat. > - The status page was updated within 10 minutes of the page. _Write yours here._ ## 07 · What was hard _Where people struggled: tools, information, timing._ Hints: - Describe the situation, not anyone's skill. - "We didn't know X" is more useful than "X was missed". Example: > - The error message pointed at the wrong system. > - The renewal steps in the runbook were 8 months old. > - It happened during the busiest hour of the week. _Write yours here._ ## 08 · Action items _Each one fixes a factor from part 05. Owner, date and ticket for every row._ Hints: - One owner per row, not a team. - Mark each as prevent, detect or recover, and check you have some of each. Example: > | Action | Type | Fixes | Owner | Due | Ticket | > | --- | --- | --- | --- | --- | --- | > | Fix auto-renewal for the new DNS provider and test a renewal in staging | Prevent | A2, A3 | Priya | 14 Mar | OPS-412 | > | Add a "What depends on this provider?" step to the migration checklist | Prevent | A2 | Priya | 28 Mar | OPS-416 | > | Alert on renewal errors and on certificates expiring within 14 days | Detect | A1, B1 | Sam | 10 Mar | OPS-413 | > | Check checkout from outside our network every minute | Detect | B2 | Sam | 14 Mar | OPS-415 | > | Show "Temporary problem, you were not charged" instead of "Payment declined" when the payments API is down | Recover | B3 | Lee | 21 Mar | SHOP-88 | > | Update the certificate runbook for the new DNS setup, with a 5-minute manual renewal | Recover | B4 | Jo | 17 Mar | OPS-414 | | Action | Type | Fixes | Owner | Due | Ticket | | --- | --- | --- | --- | --- | --- | | | | | | | | ## 09 · Thanks _Who helped, and how._ Hints: - Thank the people who noticed and spoke up, not only the fixers. Example: > Thanks to Lee in support for paging fast, to Priya for staying calm on the call, and to Jo for writing the renewal steps down while fixing them. _Write yours here._ ## 10 · Links _Dashboards, chat logs, graphs and tickets._ Hints: - Link the incident chat so readers can see the real order of events. Example: > - Incident chat: #inc-2026-03-04-checkout > - Certificate dashboard, payments API error graph _Write yours here._