Telling people.
WHY IT MATTERS
Silence is worse than bad news.
People assume the worst and contact you individually.
WHAT TO SAY FIRST
That you are aware and investigating.
WHAT TO INCLUDE
What is affected What is not When you will update again
THAT LAST POINT
Committing to an update time is more valuable than estimating a fix time.
WHERE TO PUBLISH
Somewhere independent of the affected system.
WHY
Your status page hosted on the failed server tells nobody anything.
WHAT CHANNELS TO USE
Whatever your customers actually use.
Here, that frequently means messaging as well as email.
WHAT NOT TO DO
Blame a provider publicly Give technical detail nobody needs Claim resolution before you have verified it
WHAT TO SAY WHEN RESOLVED
That it is resolved What happened, briefly What you are doing to prevent recurrence
WHAT TO AVOID IN THAT MESSAGE
Excessive apology Excuses
WHAT TO DO AFTERWARDS
Follow up with anyone significantly affected.
