Copyable template
--- UPDATE 1: INVESTIGATING --- Status: Investigating [Affected Component] Time: [Date, Time and Timezone] We are investigating an issue causing [observable user symptom, e.g., failed logins or delayed webhooks] for [scope: all users / subset of workspaces]. Affected: [Broken features]. Still working: [Unaffected systems and data safety note]. We are tracing the cause now and will post the next update by [Next Update Time and Timezone]. --- UPDATE 2: RESOLVED --- Status: Resolved - [Affected Component] Time: [Date, Time and Timezone] The issue affecting [Affected Component] is resolved as of [Resolution Time and Timezone] after we [brief fix, e.g., rolled back a migration or rerouted traffic]. Impact: Between [Start Time] and [End Time], [what failed and whether jobs were queued or retried]. Action needed: [None / specific step]. We are watching error rates closely and will share a short root-cause summary within 24 hours. --- UPDATE 3: POST-INCIDENT SUMMARY --- Subject: Incident summary: [Component] outage on [Date] Between [Start Time] and [End Time] on [Date], [Affected Component] experienced [duration] of [degraded performance / downtime]. What happened: [2 plain sentences on the technical trigger]. Impact: [Exact scope and data integrity confirmation]. Preventative changes: - [Concrete engineering fix 1] - [Concrete monitoring or safeguard fix 2] Sorry for the disruption. Reply if your account still shows stuck jobs or unexpected errors. [Your Name]
When to use this template
Post the first update within 15 minutes of confirming a customer-facing outage, even before you isolate the root cause. Users handle downtime calmly when they see you are aware and working on it rather than wondering if the issue is local.
Use all three stages below on your status page, in-app banner, or incident email thread. Commit to a next-update time until resolved, then share a brief plain-English summary within 24 hours.
How to fill each placeholder
- [observable user symptom]: Describe what customers actually see ('API requests returning 502 errors') rather than internal jargon ('pod crashloop on worker-3').
- [What is still working]: Explicitly stating that stored data is intact or that core billing continues working calms anxiety while you debug.
- [Next Update Time and Timezone]: Always commit to a real clock time (such as '14:45 UTC') and post on schedule even if the only update is 'still investigating'.
- [Action needed from you]: Tell customers plainly whether failed webhooks or uploads will replay automatically or require a manual retry.
- [What we changed so it does not repeat]: List real fixes like adding a staging migration check or queue depth alert rather than vague promises to 'improve testing'.
Filled example (made-up product)
Example using a made-up billing notification router called RelayMint: --- UPDATE 1: INVESTIGATING --- Status: Investigating delayed email and SMS alerts Time: November 12, 2026, 14:15 UTC We are currently investigating an issue causing outbound billing emails and SMS messages to queue instead of sending immediately for all RelayMint workspaces. What is affected: Outbound email and SMS delivery (delays currently around 10 minutes). What is still working: Incoming webhook ingestion is healthy, and all queued alerts are safely stored with zero message loss. We are investigating the queue worker slowdown now. Next update by 14:45 UTC. --- UPDATE 2: RESOLVED --- Status: Resolved - Delayed email and SMS alerts Time: November 12, 2026, 14:42 UTC Outbound delivery is back to normal as of 14:38 UTC after we restarted a stuck Redis consumer group and cleared the backlog. Impact summary: Between 14:04 UTC and 14:38 UTC, 1,420 alerts were delayed by up to 34 minutes. All queued messages have now been delivered. Action needed from you: None. No duplicate messages were sent. --- UPDATE 3: POST-INCIDENT SUMMARY --- Subject: Incident summary: Alert delivery delay on November 12 Between 14:04 and 14:38 UTC on November 12, RelayMint experienced a 34-minute delivery delay on outbound notifications. What happened: A worker deploy introduced an unindexed lookup on retry metadata, which slowed consumer acknowledgements until the queue backed up. Customer impact: Alerts queued during the window arrived late, with zero dropped or duplicated events. What we changed so it does not repeat: - Added the missing index and required query-plan checks in CI before worker deploys. - Lowered our queue lag pager threshold from 10 minutes to 2 minutes. Sorry for the delay in your customer notifications. Reply here if any event in your log looks off. Nadia
Mistakes to avoid
- Staying silent for two hours while hoping to fix the bug before anyone notices.
- Blaming an upstream cloud provider without explaining how you are mitigating the impact for your own users.
- Marking an incident 'Resolved' before backlogged jobs finish draining and confirming whether data was lost.
- Writing a defensive post-incident note instead of owning the trigger and listing concrete preventative changes.
Related tools and guides
Free tool
Launch Day ChecklistEverything to check before, during and after launch day, saved in your browser.Founder guide
How to Write a Changelog People Actually Read (2026)Why a public changelog is worth the effort, what goes in it, how to write entries for users, version numbers, where to publish it and how to turn it into posts.Founder guide
What to Do After Launch Day: The Two Weeks That Decide Whether It WorkedLaunch day gets the attention, but the next fourteen days decide whether a launch turns into users. A day-by-day plan for solo founders.More templates
Pritam Ghosh runs Hauntboard and maintains these templates.
Published · Updated
Ready to put your launch in front of real visitors?
A person reviews every launch on Hauntboard, and every view and visit it sends your site is published on your launch page. Free to submit.