1. LandOffer
  2. Blog
  3. All articles
  4. Customer Outage Updates: Explain Impact, Workarounds, and What Is Still Unknown

Workplace Communication

Customer Outage Updates: Explain Impact, Workarounds, and What Is Still Unknown

Write customer outage notices using confirmed impact, bounded workarounds, explicit unknowns and communication commitments. Follow a fictional progression from investigation through verified closure.

10 min readLandOffer team

A customer outage update should explain what people cannot do, what they can safely do instead, what remains unknown, and when you will update them again. Separate confirmed impact from reports still being checked. If you do not have a reliable restoration estimate, say so and give a specific next-update time rather than turning your communication schedule into a recovery promise.

This LandOffer workplace guide shows a completed fictional announcement sequence and the edits that turn internal technical notes into customer-facing messages. Every service, observation, verification, approval, and timestamp in the case is invented for illustration. No system was tested, customer contacted, or status page changed while preparing this article.

Harbor information coordinator explaining a disruption while passengers wait and a technician checks a signal

Establish the facts behind each sentence

An early notice can acknowledge confirmed symptoms while the cause is still unknown. It can still name the affected customer action, acknowledge the interruption, and set a communication commitment. Waiting for a polished cause narrative can leave customers guessing whether to retry, investigate their own setup, or stop dependent work.

Atlassian's incident communication tips recommend early acknowledgment, known-impact summaries, continuing updates, precise language, and consistency across channels. The page suggests a cadence appropriate to the situation; its example interval is not a universal requirement. Your actual incident process should establish who confirms facts and who is authorized to publish them.

For each material statement, keep its internal basis nearby: the observation, its time, its scope, and the person responsible for confirming it. “A customer reported a blank preview” supports acknowledging that report. It does not establish that every account, region, or document is affected. “The download check passed” needs a defined workflow and permission context before it becomes advice to all customers.

Keep the detailed evidence internally so the writer can check and remove unsupported claims before publication. If a fact changes, preserve the earlier update's timestamp and issue a correction rather than silently rewriting the history customers used to make decisions.

Describe impact in the customer's language

Name the task people recognize. “Document previews may remain blank” is more actionable than an internal component name. State affected interfaces or regions only as far as the evidence supports them, and distinguish an earliest confirmed failure from the actual incident start if that remains uncertain.

In the fictional case, Cedar Review is a document-review service. At 08:10 UTC, the team has confirmed that some browser document previews remain blank. The earliest confirmed occurrence is 08:02 UTC; the number of affected accounts and regional extent are still being checked. The team has not established whether every document type is affected.

Do not infer that other operations are healthy because they share no obvious dependency with preview rendering. “We are checking whether downloads are affected” is different from “downloads are unaffected.” Equally, a missing error signal is not evidence for a broad assurance about data integrity or security. Such statements require their own verified basis and the appropriate review.

Avoid percentages without a meaningful denominator and observation window. An error rate for one request path does not automatically describe the percentage of customers unable to complete their work. If you cannot state a useful measured scope, describe the confirmed behavior and the current limit of your knowledge.

Offer a workaround only within its verified limits

A workaround should be something customers can use now, with prerequisites and trade-offs explained. A proposed engineering mitigation, an untested alternate endpoint, or a feature available only to a different plan is not ready to become customer advice.

In this fictional case, by 08:35 UTC the response team has checked the existing desktop-browser Download original action for documents the signed-in user already has permission to download. That route returns the original document without requiring a new upload. It does not recreate preview-only annotations, and mobile availability has not been verified in the scenario.

The update therefore limits the workaround to that existing authorized desktop action and says what it does not provide. It does not recommend granting broader access, emailing sensitive files, or repeatedly uploading replacements. The bounded instruction lets customers assess the alternative without changing permissions or creating replacement documents.

If a workaround fails for a customer, provide the team's approved support route and useful nonsensitive context to include. Do not ask people to paste document contents, credentials, or private account details into a public incident thread. Support follow-up and public status communication serve different purposes even when they refer to the same interruption.

A completed customer announcement sequence

The following five notices are fictional drafts for one incident. All times are UTC on the same illustrative day. Jules owns customer communications; incident lead Ren confirms the operational facts before each notice is approved within the scenario. The checks described here were not performed by the article's author.

08:10 — Investigating blank document previews

“We are investigating blank document previews affecting some customers in Cedar Review's browser interface. The earliest confirmed occurrence is 08:02 UTC. We are still checking the affected account range, regions, and document types. The cause and restoration time are not confirmed. We are also checking available alternatives. Our next update will be at 08:40 UTC, or sooner if we confirm information that changes what you should do. We are sorry for the interruption to your review work.”

08:40 — Partial improvement; a limited alternative is available

“Some browser previews have begun loading again, but blank previews continue in our checks in both supported service regions. We have not confirmed the total number of affected accounts or the cause, and we do not yet have a reliable restoration estimate. On desktop browsers, users who already have download permission can use Download original to open the original document locally. This does not include preview-only annotations; we have not verified this alternative on mobile. Do not upload a replacement document to use this route. We will update you again at 09:10 UTC.”

09:10 — Preview service restored in current checks; monitoring continues

“Following a change to preview request routing at 09:02 UTC, our checks of supported preview formats in both service regions are succeeding. We are monitoring for recurring failures and reviewing remaining customer reports; this incident is still open. The download alternative remains available within the limits described above. We have not completed confirmation of the cause. Our next update will be at 09:40 UTC. If previews remain blank, please contact support through the existing in-product Help route and include the time, browser version, and document type, without sending document contents.”

09:40 — Monitoring continues; scheduled update

“Preview checks continue to succeed in both supported service regions. We have not observed new preview failures since 09:02 UTC, but we are still completing the incident closure checks and reviewing the routing change. We are keeping the incident open during that review. We do not have additional customer action to recommend. Our next update will be at 10:10 UTC, or sooner if the situation changes.”

10:10 — Preview incident resolved

“The blank-preview incident is resolved. Checks of supported preview formats in both service regions have continued to succeed, and the incident team has completed its defined closure checks. The team confirmed that an incorrect preview routing rule caused the failures and disabled that rule. The confirmed impact window was 08:02–09:02 UTC; we have not established whether failures began earlier. You can resume the normal browser preview workflow. We are sorry for the disruption. Our service owner will post a reviewed incident summary by 16:00 UTC on the next business day, covering the impact, contributing factors, and follow-up actions.”

The 08:40 notice records partial improvement without closing the incident. The 09:10 notice separates currently successful checks from ongoing monitoring and incomplete cause confirmation. The final notice relies on additional fictional closure evidence; it does not treat the passage of time as proof of recovery.

Statuspage's incident documentation defines investigating, identified, monitoring, and resolved states, with resolved requiring restored performance and elimination of the root cause. Those are that product's definitions. Match your actual tool and incident policy, and ensure its displayed state agrees with the words customers read.

A communications specialist simplifying a dense technical sheet into a clear notice while retaining uncertain details separately

Turn internal technical notes into an external update

Here is the completed editing record behind the fictional 08:40 notice. The internal note reads: “Renderer shard B requests still fail in both regions. A path change looks promising. Maybe yesterday's routing rollout? Desktop Download original succeeded on permitted documents. Mobile not checked. Preview annotations don't travel with original. Guess twenty minutes, not verified. Some previews working since 08:32. Tenant examples attached privately.”

Internal material External treatment Reason for the edit
Renderer shard B requests still fail in both regions “Blank previews continue in our checks in both supported service regions.” Translate the symptom and retain the checked scope; avoid implying every customer is affected.
Some previews working since 08:32 “Some browser previews have begun loading again.” Report improvement while keeping the continuing failure visible; improvement is not resolution.
A path change looks promising; maybe yesterday's rollout Omit the suspected cause from the announcement; state that the cause is unconfirmed Preserve the hypothesis internally without presenting chronology or suspicion as established cause.
Desktop Download original succeeded on permitted documents; mobile not checked Limit the alternative to existing desktop download permission; explicitly retain the mobile uncertainty Make eligibility and evidence limits part of the instruction.
Preview annotations do not travel with original Explain that preview-only annotations are excluded Customers need this limitation before choosing the alternative.
Guess twenty minutes, not verified Remove the restoration estimate; commit to the 09:10 update A responder's guess is not an approved recovery forecast.
Private tenant examples Keep in restricted incident evidence; do not include identities in the public text Customer-specific evidence need not become public disclosure.

The editing record preserves the difference between observation, hypothesis, and an actionable instruction. A possible cause becomes an explicit unknown, a narrow successful check becomes a bounded alternative, and an optimistic guess becomes a communication deadline the team can own.

Before approving the draft, Ren confirms that the described checks and remaining impact match the current incident record. Jules verifies the 09:10 UTC update commitment and the desktop, permission, annotation, and mobile limits in the instruction. Their fictional approval concerns this notice only; it is not evidence of any actual authorization or incident response.

Keep the next update separate from a restoration estimate

An update time is a promise to communicate. A restoration estimate is a forecast about service behavior. If an approved estimate exists, describe its basis and uncertainty according to your team's process, and revise it when that basis changes. Do not supply an estimate simply because the template has a blank for one.

Atlassian's communication examples distinguish potential interruptions from full and partial outages, and include next-update commitments in outage examples. Replace their placeholders with verified incident facts and use your established publication authorization.

When the scheduled time arrives without new findings, say what is still happening and repeat the next commitment. The fictional 09:40 notice does this without inventing progress. If a promised update is late, acknowledge the missed communication time and provide the current verified status; do not backdate the notice to make the sequence look tidy.

Close the customer loop without overstating certainty

Keep the approved incident entry as the shared reference for support replies and other customer channels. Shorter notices can link to that entry, but should not contradict its impact, workaround limits, or current state. Preserve timestamp and timezone when copying an update so an old message is not mistaken for the latest status.

Before a final notice, check that recovery covers the customer workflow that failed and that any remaining impact is disclosed. A restored new-request path may coexist with delayed older work. In that situation, describe both and retain the incident or a clearly linked follow-up according to your process; do not hide unfinished recovery behind a green label.

For your next incident draft, verify each impact statement, test or confirm the exact proposed alternative through the authorized response team, and give the next communication owner a realistic commitment. After restoration, retain the timestamped notices and their evidence, then deliver the reviewed follow-up at the time your team committed to provide it. Continue with other workplace guides through LandOffer.

Sources and Further Reading